Plasminogen activator inhibitor type-1 deficiency attenuates murine antigen-induced arthritis.

OBJECTIVE: To examine the role of plasminogen activator inhibitor type-1 (PAI-1), the major fibrinolytic inhibitor, in vivo during murine antigen-induced arthritis (AIA). METHODS: AIA was induced in PAI-1-deficient mice and control wild-type mice. Arthritis severity was evaluated by technetium 99m (99mTc) uptake in the knee joints and by histological scoring. Intra-articular fibrin deposition was examined by immunohistochemistry and synovial fibrinolysis quantitated by tissue D-dimer measurements and zymograms. RESULTS: Joint inflammation, quantitated by 99mTc uptake, was significantly reduced in PAI-1(-/-) mice on day 7 after arthritis onset (P<0.01). Likewise, synovial inflammation, evaluated by histological scoring, was significantly decreased in PAI-1-deficient mice on day 10 after arthritis onset (P<0.001). Articular cartilage damage was significantly decreased in PAI-1(-/-) mice, as shown by histological grading of safranin-O staining on day 10 after arthritis onset (P<0.005). Significantly decreased synovial accumulation of fibrin was observed by day 10 in arthritic joints of PAI-1(-/-) mice (P<0.005). Accordingly, the synovial tissue content of D-dimers, the specific fibrin degradation products generated by plasmin, were increased in PAI-1(-/-) mice (P<0.02). Finally, as expected, PA activity was increased in synovial tissues from PAI-1(-/-) mice, as shown by zymographic analysis. CONCLUSIONS: These results indicate that deficiency of PAI-1 results in increased synovial fibrinolysis, leading to reduced fibrin accumulation in arthritic joints and reduced severity of AIA.

Effect of thrombin inhibition on synovial inflammation in antigen induced arthritis.

OBJECTIVE: To determine the effect of the thrombin inhibitor, hirudin, on the pathogenesis of murine antigen induced arthritis (AIA). METHODS: AIA was induced by intra-articular injection of methylated bovine serum albumin in the knee joints of previously immunised mice. Hirudin (injected subcutaneously 3 x 200 microg/mouse/day) was given over 13 days, starting three days before arthritis onset, and its anticoagulant effect monitored by clotting times. Arthritis severity was evaluated by technetium-99m ((99m)Tc) uptake in the knee joints and by histological scoring. In addition, intra-articular fibrin deposition was examined by immunohistochemistry, and synovial cytokine mRNA expression measured by RNase protection. RESULTS: Joint inflammation, measured by (99m)Tc uptake, was significantly reduced in hirudin treated mice at days 7 and 10 after arthritis onset. Histologically, synovial thickness was markedly decreased in hirudin treated mice compared with untreated ones. By contrast, no difference in articular cartilage proteoglycan content was found between both groups. Intra-articular fibrin deposition and synovial interleukin 1beta mRNA levels, were slightly reduced ( approximately 20%) in arthritic joints from hirudin treated mice compared with untreated ones at day 10 of AIA. CONCLUSION: Hirudin reduces joint inflammation associated with AIA by fibrin-dependent and independent mechanisms.

Identification of amino acid residues essential for high aflatoxin B1-8,9-epoxide conjugation activity in alpha class glutathione S-transferases through site-directed mutagenesis.

Mice constitutively express glutathione S-transferase mGSTA3-3 in liver. This isoform possesses uniquely high conjugating activity toward aflatoxin B1-8,9-epoxide (AFBO), thereby protecting mice from aflatoxin B1-induced hepatocarcinogenicity. In contrast, rats constitutively express a closely related GST isoenzyme, rGSTA3-3, with low AFBO activity and, therefore, are sensitive to aflatoxin B1 exposure. Although the two GSTs share 86% sequence identity and have similar catalytic activities toward 1-chloro-2,4-dinitrobenzene (CDNB), they have an approximately 1000-fold difference in catalytic activity toward AFBO. To identify amino acids that confer high activity toward AFBO, non-conserved rGSTA3-3 residues were replaced with mGSTA3-3 residues in two regions believed to form the substrate binding site. Twenty-one mutant rGSTA3-3 enzymes were generated by site-directed mutagenesis using combinations of nine different residues. Except for the E208D mutant, single mutations of rGSTA3-3 produced enzymes with no detectable AFBO activity. Generally, AFBO conjugation activity increased in additive fashion as mGSTA3-3 residues were introduced into the rGSTA3-3 enzyme with the six site mutant E104I/H108Y/Y111H/L207F/E208D/V217K displaying the highest AFBO activity (40 nmol/mg/min) of all the mutant enzymes. When this mutant enzyme was further modified by three additional substitutions (D103E/I105M/V106I) AFBO conjugation activity decreased 14-fold to 2. 8 nmol/mg/min. Although wild-type mGSTA3-3 AFBO conjugation activity (265 nmol/mg/min) could not be obtained by our rGSTA3-3 mutants, we were able to identify six mGSTA3-3 residues; Ile104, Tyr108, His111, Phe207, Asp208, and Lys217 that, when collectively substituted into rGSTA3-3, substantially increased (>200-fold) glutathione conjugation activity toward AFBO.

[Hemodynamic effects of maximum laser transmyocardial revascularization]. / Effets hémodynamiques de la revascularisation transmyocardique au laser maximale.

The aim of this study was to determine the haemodynamic reaction of a heart in which the left lateral wall was bored with laser channels until cardio-vascular collapse. Four calves were selected for the study. Series of five channels were bored with a 1.75 mm diameter laser probe. Each series was followed by a break of three minutes at the end of which haemodynamic parameters were recorded. The evolution of these parameters then underwent linear regression analysis. Calculations were made according to the percentage of the total number of channels, in order to standardize the results between the animals. Respectively 150, 155, 270 and 285 channels were bored. In each case, all haemodynamic parameters dropped abruptly during the last series of channels. Central venous pressure, mean pulmonary artery pressure, wedge pressure, mean arterial pressure and cardiac output followed a linear regression slope which did not significantly differ from zero. Heart rate only increased progressively with a liner regression slope significantly different from zero. In this acute model, haemodynamic parameters, except heart rate, did not correlate to the extent of damage imposed on the left ventricle. Thus, the presence of haemodynamic stability does not exclude important myocardial damage in an acute situation. This can be found in a clinical setting.

Exacerbation of antigen-induced arthritis in urokinase-deficient mice.

In rheumatoid arthritis, synovial expression of urokinase (uPA) activity is greatly increased (Busso, N., V. Péclat, A. So, and A. -P. Sappino. 1997. Ann. Rheum. Dis. 56:550- 557). We report the same effect in murine antigen-induced arthritis. uPA-mediated plasminogen activation in arthritic joints may have deleterious effects via degradation of cartilage and bone matrix proteins as well as beneficial effects via fibrin degradation. We evaluated these contrasting effects in vivo by analyzing the phenotype of uPA-deficient (uPA-/-) and control mice during antigen-induced arthritis. Joint inflammation was comparable in both groups up to day 3 and subsequently declined in control mice, remaining significantly elevated in uPA-/- mice on days 10 and 30 after arthritis onset. Likewise, synovial thickness was markedly increased in uPA-deficient mice persisting for up to 2 mo, whereas it subsided in control animals. Bone erosion was exacerbated in uPA-/- mice on day 30. By contrast, no difference in articular cartilage proteoglycan content was found between both groups. Significantly increased accumulation of fibrin was observed by day 30 in arthritic joints of uPA-/- mice. We hypothesized that synovial fibrin deposition plays a role in joint inflammation. Accordingly, defibrinogenation of uPA-/- mice by ancrod significantly decreased the sustained joint inflammation. All the above observations were reproducible in plasminogen-deficient (Pln-/-) mice. In conclusion, synovial fibrin deposition plays a role as a nonimmunological mechanism which sustains chronic arthritis.

Air trapping ability of the Spiral Gold membrane oxygenator: an ex vivo study.

Despite an overall improvement in cardiopulmonary bypass (CPB) technology and materials, air emboli still occur. The latest generation membrane oxygenator from Bentley Laboratories, the SpiralGold, was tested ex vivo for its air handling ability. The study was conducted on four calves. Bolus amounts of air of 10, 15 and 20 cm3 were each injected three times, upstream of the oxygenator and a bubble detector located directly downstream. The amount of bubbles was measured semiquantitatively on a 10 unit scale (U one semiquantitative unit). The animals were killed 10 days after the CPB. When 10 cm3 of air was injected, no bubbles were detected. With 15 and 20 cm3, respectively, 1 +/- 1.5 and 5 +/- 3.3 U of bubbles were detected. Despite a total of 135 cm3 of air injected as large bolus amounts, all the animals survived without any obvious neurological deficit secondary to air bubble manipulation. In conclusion, the SpiralGold oxygenator per se can reliably trap an air bolus of up to 10 cm3. This feature should be taken into account when choosing an oxygenator, as it offers an additional barrier to air bubbles in the CPB circuit.

Oleate uptake by isolated hepatocytes and the perfused rat liver is competitively inhibited by palmitate.

Competition for uptake between long-chain free fatty acids has been difficult to document, because there has been no algorithm for computing unbound concentrations of two fatty acids simultaneously in solution with albumin. We modified an iterative procedure to permit this computation and studied initial [3H]oleate uptake by isolated hepatocytes and steady-state uptake by the single-pass perfused rat liver from 600 microM bovine serum albumin solutions containing various concentrations of oleate in the presence and absence of palmitate. In both systems, the Michaelis-Menten constant was significantly higher in the presence of palmitate than in its absence, whereas the maximal reaction velocity was unaltered, indicating competitive inhibition. In additional experiments employing the multiple transhepatic indicator-dilution technique, the influx rate constant and permeability-surface area product for oleate influx were significantly reduced by palmitate, confirming that the competition observed in the conventional perfused liver studies was at the influx step. Long-chain fatty acid uptake has now been shown to exhibit all the kinetic properties of facilitated transport and cannot be attributed solely to passive diffusion.

Dibutyryl cAMP reduces nonparenchymal cell damage during cold preservation of rat livers.

Sinusoidal lining cells are the main target for cold preservation injury and are further damaged with reperfusion. Different agents known to increase intracellular cAMP levels have been shown beneficial. This study was designed to assess the possible protective effect of a cAMP analogue on nonparenchymal cells of rat livers, during cold storage and during reperfusion. Parameters reflecting the status of the liver microvasculature were analyzed. The initial effluent collected after preservation reflects release during the cold storage period; therefore we measured interleukin-1 (IL-1) and endothelin-1 (ET-1) levels in these samples in order to detect and quantitate the degree of activation and/or disruption of Kupffer and sinusoidal endothelial lining cells. Rat livers were harvested after in situ flush with Ringer's lactate with or without 2 mM dibutyryl cAMP, excised, and stored in the same solution at 4 degrees C. After 6 hr, livers were perfused with Krebs-Henseleit buffer for 90 min. Physiological parameters were monitored throughout the perfusion. Perfusate samples were collected every 30 min for RIA measurements of IL-1 and ET-1. Treatment resulted in a significant decrease in release of ET-1 and IL-1 during storage. Likewise, livers treated with cAMP had a significantly improved bile output and decreased portal vein resistance during reperfusion. The beneficial effect granted by the analogue during cold storage and reperfusion was evident on parenchymal and nonparenchymal cells. Levels of ET-1 and IL-1 in the caval effluent confirm and quantitate preservation damage.

The hepatocellular uptake of free fatty acids is selectively preserved during starvation.

BACKGROUND/AIMS: The liver loses protein during fasting. This study sought to determine if hepatic protein loss during fasting selectively preserves functions important to survival such as uptake of fatty acids, which are major energy substrates in that condition. METHODS: Initial [3H]oleate uptake and efflux rates in hepatocytes from starved (for 48 hours) and fed male rats were measured in media containing 250 mumol/L albumin at oleate/albumin ratios of 0.2:1-2:1. Uptake rates of sulfobromophthalein, taurocholate, and glucose were also determined. RESULTS: Initial oleate uptake rate was saturable with respect to unbound oleate concentration. Maximum initial velocity expressed per cell number did not differ between fasted and fed animals, but measured cell volume and estimated surface area were decreased in starved vs. fed hepatocytes (921 +/- 21 vs. 1623 +/- 58 microns2, respectively; P < 0.001). Consequently, when expressed per surface area, maximum initial velocity was greater in starved cells (17 +/- 3 vs. 10 +/- 2 [pmol.min-1.micron2] x 10(-7); P < 0.02). Expressed similarly, oleate efflux was also greater from starved hepatocytes and was inhibited by an antibody to plasma membrane fatty acid binding protein (FABPpm). FABPpm concentration per unit area of plasma membrane also increased in starved hepatocytes (P < 0.05). By contrast, uptake rates of sulfobromophthalein, taurocholate, and glucose by starved hepatocytes were decreased when expressed per cell number and unchanged per unit area. CONCLUSIONS: During fasting, the hepatocellular uptake mechanism for oleate is selectively preserved compared with those for sulfobromophthalein, taurocholate, or glucose.

Oleate uptake kinetics in the perfused rat liver are consistent with pseudofacilitation by albumin.

We measured uptake of a representative free fatty acid, oleate, by the single-pass perfused rat liver at oleate:albumin molar ratios of 0.01 to 2:1. For each ratio, uptake was studied at albumin concentrations from 50 to 600 microM. When uptake velocity was plotted as a function of the albumin concentration, the data at each ratio exhibited a pseudosaturation pattern as previously observed in isolated cells (J Clin Invest 84: 1325). At a physiologic albumin concentration of 600 microM, a plot of uptake vs. unbound oleate concentrations was best fitted by the Michaelis-Menten equation (Vmax = 235 +/- 8.8 nmol.min-1.g.liver-1; Km = 130 +/- 12 nM). As the albumin concentration was increased from 50 to 250 microM, the unbound oleate clearance, calculated by either the undistributed sinusoidal or venous equilibrium models, increased progressively, in violation of conventional pharmacokinetic theory, indicating an enhancing effect of albumin on ligand uptake at low albumin concentrations. In contrast, there was no significant difference between measures of unbound clearance at albumin concentrations of 350 and 600 microM. To explain this phenomenon, the clearance data were examined for evidence of facilitation (accelerated dissociation of ligand:albumin complexes) by the clearance ratio test ("square root rule"). All deviations from the predictions of conventional theory were entirely attributable to pseudofacilitation. No data required explanation by a true facilitation model.

Enzymatic characteristics of chimeric mYc/rYc1 glutathione S-transferases.

Mice are resistant to aflatoxin carcinogenicity primarily due to expression of a glutathione S-transferase (mYc) with high catalytic activity toward aflatoxin B1-8,9-epoxide (AFBO). In contrast, rats are more sensitive to aflatoxin carcinogenicity due to the constitutive expression of a glutathione S-transferase with relatively low catalytic activity toward AFBO (rYc1). To identify the contribution of different regions of the mYc protein that confer high catalytic activity toward AFBO, six chimeric mYc/rYc1 GST enzymes were generated utilizing full and partial restriction enzyme digestions at two conserved StyI sites in the mYc and rYc1 complementary DNAs (between amino acid residues 56-57 and 142-143). Recombinant wild-type and chimeric glutathione S-transferases were bacterially expressed, affinity purified, and their catalytic activities measured toward AFBO, delta 5-androstene-3,17-dione, 1-chloro-2,4-dinitrobenzene, and ethacrynic acid. The set of chimeras displayed a wide range of catalytic activities toward the substrates assayed. The chimeras with the greatest activity toward AFBO were 1:56rat-57: 221mouse and 1:56mouse-57:142rat-143:221mouse, with AFBO conjugating activities 200 and 8 times greater than wild-type rYc1, respectively. These results demonstrate that the residues that confer high AFBO conjugation activity in mYc are located in the region spanning residues 57-221.

Unbound ligand drives hepatocyte taurocholate and BSP uptake at physiological albumin concentration.

We have recently shown (D. Sorrentino, R.B. Robinson, C.-L. Kiang, and P.D. Berk. J. Clin. Invest. 84: 1325-1333, 1989) that, in a variety of isolated cell types, the uptake of oleate at physiological albumin concentrations is consistent with traditional pharmacokinetic theory (i.e., driven by unbound ligand). Lower albumin concentrations were associated with a deviant uptake pattern for which alternative theories have been proposed. Whether other classes of organic anions exhibit similar behavior is unknown. Therefore, we examined the effect of albumin on uptake of two widely studied organic anions, sulfobromophthalein (BSP) and taurocholate. Initial uptake velocity of [35S]BSP and [3H]taurocholate by isolated hepatocytes was studied employing a fixed albumin concentration and ligand-to-albumin molar ratios from 0.01:1 to 2:1 for taurocholate and 0.031:1 to 0.75:1 for BSP. In other experiments, albumin and ligand were altered in parallel, keeping their molar ratio constant. Unbound taurocholate concentrations were measured directly by equilibrium dialysis; unbound BSP concentrations were calculated from published data (K.J. Baker and S.E. Bradley. J. Clin. Invest. 45: 281-287, 1966). At 600 microM albumin, uptake of both ligands was a function of the unbound ligand concentration. At low ligand-to-albumin molar ratios and consequent unbound ligand concentrations this relationship was linear; over the entire range of unbound ligand concentrations studied, both ligands exhibited Michaelis-Menten kinetics, with definable maximal velocity and Michaelis constant values. At low albumin concentrations, the relationships between uptake and unbound ligand were unchanged for taurocholate; however, BSP exhibited altered kinetics similar to those observed with oleate. Nontraditional uptake kinetics at low albumin concentrations appear to correlate with very high affinity for albumin.

Purification of interleukin-1 beta converting enzyme, the protease that cleaves the interleukin-1 beta precursor.

We have purified the IL-1 beta converting enzyme from the THP-1 cell line using standard chromatographic techniques and obtained the N-terminal amino acid sequence of this novel protein. After stimulation of THP-1 cells with lipopolysaccharide, hydroxyurea, and silica, the protease was solubilized by multiple freeze/thawing. The protein was purified by ion-exchange chromatography, affinity chromatography on blue agarose, gel filtration, and chromatofocusing. The molecular weight of the protein is approximately 22,000 Da and the pI is between 7.1 and 6.8. The overall yield for this procedure was 16% of the activity found in the initial cell lysates. An antiserum raised against a peptide based on the N-terminus was used to precipitate the protease, confirming our identification of the 22,000-Da protein as the IL-1 beta converting enzyme.

Molecular cloning of the interleukin-1 beta converting enzyme.

Interleukin-1 beta (IL-1 beta) mediates a wide range of immune and inflammatory responses. The active cytokine is generated by proteolytic cleavage of an inactive precursor. A complementary DNA encoding a protease that carries out this cleavage has been cloned. Recombinant expression in COS-7 cells enabled the cells to process precursor IL-1 beta to the mature form. Sequence analysis indicated that the enzyme itself may undergo proteolytic processing. The gene encoding the protease was mapped to chromosomal band 11q23, a site frequently involved in rearrangement in human cancers.

Hepatocellular 22Na+ uptake: effect of oleate.

Whether cellular oleate uptake is Na+ coupled remains controversial. Our present studies document that hepatocellular [3H]oleate uptake is unaltered by isomotic substitution of Na+ with K+, Li+, or sucrose in Hanks' HEPES buffer. In parallel studies 22Na+ uptake was significantly (P less than 0.02) increased by concentrations of alanine that yielded [3H]-alanine uptakes greater than 24% of basal 22Na+ uptake when both were expressed as nanomoles per minute per 10(6) hepatocytes. Although [3H]glutamine uptake exceeded this threshold, maximal specific [3H]taurocholate uptake did not. Consistent with the observations with alanine, addition of glutamine, but not taurocholate, to the incubations resulted in a significant increase in 22Na+ uptake. The ionophore monensin increased uptake of 22Na+ under all conditions. Both in the absence and in the presence of HCO3- (4 and 25 mM), specific [3H]oleate uptake was sufficient to elicit a readily detectable effect on 22Na+ uptake, if Na+ and oleate were cotransported. However, addition of 1 mM oleate did not affect 22Na+ uptake in the absence of HCO3- (5.4 +/- 0.6 vs. control 5.4 +/- 0.4 nmol.min-1.10(6) cells-1) as well as in the presence of 4mM HCO3- (10.4 +/- 1.8 vs. control 9.9 +/- 1.1) and 25 mM HCO3- (10 +/- 1.3 vs. control 10.8 +/- 0.5). These data indicate that the predominant component of hepatocellular oleate uptake is not directly associated with Na+ influx. They do not exclude, however, a more complex or indirect link.

Functional and morphological features of isolated hepatocytes preserved in University of Wisconsin solution.

Although several studies suggest that hepatic graft failure after cold ischemia results from nonparenchymal cell damage, other data indicate that hepatocellular ATP content is significantly correlated with the transplantation success rate. In this study, we have conducted a systematic investigation of various aspects of cell viability and function of isolated hepatocytes stored at 4 degrees C for 24 and 48 hr in either University of Wisconsin solution or Hanks' HEPES buffer, a control solution clinically unsuitable for organ preservation. After 24 hr, hepatocytes stored in Hanks' HEPES buffer had viability (measured by trypan blue exclusion and ALT and lactic dehydrogenase leakage), transport function (measured by 22Na+ and [3H]taurocholate uptake) and cell size similar or only slightly altered when compared with freshly isolated and University of Wisconsin solution-stored hepatocytes. ATP content was decreased in both groups; however, the reduction was much greater in Hanks' HEPES buffer-stored cells. Furthermore, ATP regenerating capacity was greatly reduced in Hanks' HEPES buffer- stored but not in University of Wisconsin solution-stored hepatocytes. By 48 hr viability and function of Hanks' HEPES buffer-stored hepatocytes were decreased; University of Wisconsin solution afforded partial protection. When examined by light and electron microscopy, cells stored in both University of Wisconsin solution and Hanks' HEPES buffer for 24 hr appeared essentially normal except for the presence of numerous membrane blebs in the Hanks' HEPES buffer group. Tissue sections of livers preserved in Hanks' HEPES buffer but not in University of Wisconsin solution revealed the presence of extensive amounts of blebs in the sinusoidal lumen and loss of endothelial elements.(ABSTRACT TRUNCATED AT 250 WORDS)

Characterization of a membrane antigen of Leishmania amazonensis that stimulates human immune responses.

To investigate human immune responses to defined leishmania Ag we have begun to characterize biochemically and immunologically, an abundant 42-kDa surface Ag of Leishmania amazonensis, a causative agent of human leishmaniasis. We have shown that this Ag, La gp42, is expressed on the surface of L. amazonensis promastigotes, being anchored to the membrane by a glycosyl-phosphatidylinositol moiety. As demonstrated by lectin blotting studies, La gp42 is glycosylated, binding both Con A and wheat germ agglutinin. Immunologically, La gp42 is strongly recognized by sera from patients with different forms of leishmaniasis as well as by patients with Chagas' disease. In addition, we show that purified La gp42 stimulates the proliferation of human T lymphocytes obtained from several leishmaniasis patients. Finally, the N-terminal sequence of La gp42 was obtained and a serologically cross-reactive 42-kDa protein with a homologous sequence was identified in Leishmania major.

Identification of a ligand for the c-kit proto-oncogene.

We report the purification and N-terminal amino acid sequence of a novel mast cell growth factor, termed MGF, from the supernatants of a murine stromal cell line. A panel of interleukin 3-dependent cell lines were screened for responsiveness to partially purified MGF in [3H]thymidine incorporation assays; proliferative stimulation of these cells in response to MGF correlated with expression of mRNA for the c-kit protooncogene. MGF was shown to be a ligand for c-kit by cross-linking 125I-labeled MGF to c-kit-expressing cells with subsequent immunoprecipitation of the complex with antiserum specific for the C-terminus of c-kit. This establishes MGF as a ligand for the c-kit protein.

Collagen cross-linking in human bone and articular cartilage. Age-related changes in the content of mature hydroxypyridinium residues.

The concentration in collagen of hydroxypyridinium cross-linking amino acids was measured in samples of bone and cartilage from human subjects aged from 1 month to 80 years. Cortical and cancellous bone samples were dissected and analysed separately. In both bone and cartilage, the content of this mature form of cross-link reached a maximum by 10-15 years of age (the amount in cartilage being 5-10 times that in bone), then stayed essentially in the same range throughout adult life. In bone the ratio of the two chemical variants of the mature cross-link, hydroxylysylpyridinoline to lysylpyridinoline, was constant throughout adult life at 3.5:1, whereas in cartilage it was always greater than 10:1. The ratio of hydroxypyridinium cross-links to borohydride-reducible keto-amine cross-links also changed with age. The reducible cross-links in bone collagen decreased steeply in content between birth and 25 years, but persisted in significant amounts throughout adult life. Reducible cross-links had virtually disappeared from cartilage by 10-15 years of age, being replaced by hydroxypyridinium residues, their maturation products. Cancellous and cortical bone collagens showed similar trends with age in their content of mature cross-links, though for each subject the concentration in cancellous bone was always lower than in cortical bone, presumably reflecting the higher turnover rate and hence the more immature state of cancellous bone.