On extracting design principles from biology: I. Method-General answers to high-level design questions for bioinspired robots.

When millions of years of evolution suggest a particular design solution, we may be tempted to abandon traditional design methods and copy the biological example. However, biological solutions do not often translate directly into the engineering domain, and even when they do, copying eliminates the opportunity to improve. A better approach is to extract design principles relevant to the task of interest, incorporate them in engineering designs, and vet these candidates against others. This paper presents the first general framework for determining whether biologically inspired relationships between design input variables and output objectives and constraints are applicable to a variety of engineering systems. Using optimization and statistics to generalize the results beyond a particular system, the framework overcomes shortcomings observed of ad hoc methods, particularly those used in the challenging study of legged locomotion. The utility of the framework is demonstrated in a case study of the relative running efficiency of rotary-kneed and telescoping-legged robots.

On extracting design principles from biology: II. Case study-the effect of knee direction on bipedal robot running efficiency.

Comparing the leg of an ostrich to that of a human suggests an important question to legged robot designers: should a robot's leg joint bend in the direction of running ('forwards') or opposite ('backwards')? Biological studies cannot answer this question for engineers due to significant differences between the biological and engineering domains. Instead, we investigated the inherent effect of joint bending direction on bipedal robot running efficiency by comparing energetically optimal gaits of a wide variety of robot designs sampled at random from a design space. We found that the great majority of robot designs have several locally optimal gaits with the knee bending backwards that are more efficient than the most efficient gait with the knee bending forwards. The most efficient backwards gaits do not exhibit lower touchdown losses than the most efficient forward gaits; rather, the improved efficiency of backwards gaits stems from lower torque and reduced motion at the hip. The reduced hip use of backwards gaits is enabled by the ability of the backwards knee, acting alone, to (1) propel the robot upwards and forwards simultaneously and (2) lift and protract the foot simultaneously. In the absence of other information, designers interested in building efficient bipedal robots with two-segment legs driven by electric motors should design the knee to bend backwards rather than forwards. Compared to common practices for choosing robot knee direction, application of this principle would have a strong tendency to improve robot efficiency and save design resources.

Mice over-expressing salmon calcitonin have strongly attenuated osteoarthritic histopathological changes after destabilization of the medial meniscus.

OBJECTIVE: Calcitonin is well-known for its inhibitory actions on bone-resorbing osteoclasts and recently potential beneficial effects on cartilage were shown. We investigated effects of salmon calcitonin (sCT) on the articular cartilage and bone, after destabilization of the medial meniscus (DMM) in normal and sCT over-expressing mice. DESIGN: Bone phenotype of transgenic (TG) C57Bl/6 mice over-expressing sCT at 6 months and 12 months was investigated by (1) serum osteocalcin and urinary deoxypyridinoline and (2) dynamic and normal histomorphometry of vertebrae bodies. In subsequent evaluation of cartilage and subchondral bone changes, 44 10-week old TG or wild-type (WT) mice were randomized into four groups and subjected to DMM or sham-operations. After 7 weeks animals were sacrificed, and knee joints were isolated for histological analysis. RESULTS: Trabecular bone volume (BV/TV) increased 150% after 6 months and 300% after 12 months in sCT-expressing mice when compared to WT controls (P<0.05). Osteoblast number, bone formation rate and osteocalcin measurements were not affected in TG mice over-expressing sCT. In WT animals, a 5-fold increase in the quantitative erosion index was observed after DMM, and the semi-quantitative OARSI score showed over 400% (P<0.001) increase, compared to sham-operated WT mice. DMM-operated TG mice were protected against cartilage erosion and showed a 65% and 64% (P<0.001) reduction, respectively, for the two histopathological evaluation methods. CONCLUSIONS: sCT over-expressing mice had higher bone volume, and were protected against cartilage erosion. These data suggest that increased levels of sCT may hamper the pathogenesis of osteoarthritis (OA). However more studies are necessary to confirm these preliminary results.

Osteopenic mice: animal models of the aging skeleton.

While our understanding of the developmental biology of the skeleton, like that of virtually every other subject in biology, has been transformed by recent advances in human and mouse genetics, we still know very little, in molecular and genetic terms, about skeletal physiology. Thus, among the many questions that are largely unexplained are the following: why is osteoporosis mainly a women's disease? How is bone mass maintained nearly constant between the end of puberty and the arrest of gonadal functions? Molecular genetics has emerged as a powerful tool to study previously unexplored aspects of the physiology of the skeleton. Among mammals, mice are the most promising animals for this experimental work. This has been previously demonstrated e.g. through the tremendous impact of the different osteopetrotic models on our molecular understanding of osteoclastic bone resorption. Until recently the only way of studying bone loss situations and osteoporosis in mice was by using ovariectomy with all its limitations. Today, however, we have access to more sophisticated osteoporotic mouse-models from four different origins: Transgenic mice (HSV-TK), knock-out mice (OPG), inbred-strains (SAMP6), and through physiological modulation (icv application). These new models have already taught us several important lessons. The first is, that bone remodeling is more than just an autocrine/paracrine process. Multiple experimental evidence has demonstrated that the latter regulation exists, but genetics prove that there is no functional cross-control between resorption and formation. The second lesson is, that remodeling is, at least in part, subject to central regulation. Thus, osteoporosis is partly a central or hypothalamic disease. However, the most dramatic change and the most important advantage we feel is, that today we have models to test a new hypothesis regarding the etiology of osteoporosis before it turns to dogma. Taken together, mouse-studies may lead to a shift in our physiological understanding of skeleton biology and to the emergence of novel paradigms. These, in turn, should help us to devise new treatments for degenerative diseases of the skeleton such as osteoporosis and its associated clinical problems.

[Leptin: factor in the central nervous system regulation of bone mass. Development of a new understanding of bone remodeling, skeletal reconstruction, skeletal preservation and skeletal repair]. / Leptin: Faktor in der zentralnervösen Regulation der Knochenmasse. Entwicklung eines neuen Verständnisses von Knochenremodeling, Skelettumbau, Skeletterhaltung und Skelettreparatur.

Bone remodeling is the physiologic process used by vertebrates to maintain a constant bone mass between the end of puberty and gonadal failure. Besides the well-characterized and critical local regulation of bone remodeling, recent genetic studies have shown that there is a central control of bone formation, one aspect of bone remodeling. This central regulation involves leptin, an adipocyte-secreted hormone that controls body weight, reproduction, and bone remodeling following binding to its receptor located on the hypothalamic nuclei. This genetic result in rodents is in line with clinical observations in humans and offers a whole new direction for research in bone physiology.

Sequestration of aggregated LDL by macrophages studied with freeze-etch electron microscopy.

The detailed morphology of macrophages involved in the uptake and intracellular processing of low density lipoprotein (LDL) and, ultimately, formation of macrophage-derived foam cells of atherosclerotic lesions has long fascinated investigators. This study examined localization of LDL in subcellular compartments of macrophage-derived intimal foam cells in cardiac valves isolated from rabbits by diet-induced hypercholesterolemia and, as an in vitro model of formation of foam cells, in cultured human monocyte-macrophages incubated for 2;-120 h with aggregated LDL produced by vortexing or phospholipase C lipolysis. The quasi-three-dimensional morphology of macrophages involved in endocytosis was preserved by ultrarapid freezing and freeze-etch microscopy in conjunction with thin-section electron microscopy. This approach produced unique images of subcellular compartments in human monocyte-macrophages involved in the uptake and processing of aggregated LDL with a clarity not previously reported. Three-dimensional ultrastructural analyses revealed a complex network of coated and uncoated vesicles, surface-connected saclike compartments, and endosomal/lysosomal compartments including the labyrinth of vesicular/tubular lysosomes all enmeshed in the microtubular, microfilament cytoskeletal network. These dynamic views of subcellular structures at the high resolution of the electron microscope provide an additional framework to better understand how lipoprotein particles are transported into, and processed within, macrophages during foam cell formation in atherogenesis.

Transgenic and knock out mice in skeletal research. Towards a molecular understanding of the mammalian skeleton.

Our understanding of the biology of the skeleton, like that of virtually every other subject in biology, has been transformed by recent advances in human and mouse genetics. Among mammals, mice are the most promising animals for this experimental work. Because extensive genetic information exists, many mouse mutations are known, and cells from early mouse developmental stages are accessible, scientists have developed transgenic mice - mice in which a gene is introduced or ablated in the germ line. Thus far, we have analyzed more than 100 different transgenic and knock out models with various skeletal phenotypes, covering the major aspects of both skeletal development and skeletal maintenance. Based on these results we here present a first perspective on transgenic and gene knock out animals in skeletal research, including insights in signaling pathways controlling endochondral bone formation, in the regulation of osteoblast function, osteoclastic bone resorption and in bone tumorigenesis, as well as the central control of bone formation. The use of transgenic mice to dissect and analyze regulatory mechanisms in bone cell physiology and the pathogenesis of human bone diseases is an extremely powerful experimental tool. The data presented here demonstrate that the successful convergence of novel genetic approaches with the established and fundamental knowledge of bone biology has made a beginning.

Endothelial responses to oxidized lipoproteins determine genetic susceptibility to atherosclerosis in mice.

BACKGROUND: Oxidized LDL has been found within the subendothelial space, and it exhibits numerous atherogenic properties, including induction of inflammatory genes. We examined the possibility that variations in endothelial response to minimally modified LDL (MM-LDL) constitute one of the genetic components in atherosclerosis. METHODS AND RESULTS: By a novel explant technique, endothelial cells (ECs) were isolated from the aorta of inbred mouse strains with different susceptibilities to diet-induced atherosclerosis. Responses to MM-LDL were evaluated by examining the expression of inflammatory genes involved in atherosclerosis, including monocyte chemotactic protein-1 (MCP-1) and macrophage-colony-stimulating factor (M-CSF), an oxidative stress gene, heme oxygenase-1 (HO-1), and other, noninflammatory, genes. ECs from the susceptible mouse strain C57BL/6J exhibited dramatic induction of MCP-1, M-CSF, and HO-1, whereas ECs from the resistant strain C3H/HeJ showed little or no induction. In contrast, ECs from the 2 strains responded similarly to lipopolysaccharide. CONCLUSIONS: These data provide strong evidence that genetic factors in atherosclerosis act at the level of the vessel wall.

Variation within the glycoprotein B gene of human cytomegalovirus is due to homologous recombination.

Human cytomegalovirus (HCMV) strains can be classified into different glycoprotein B (gB) genotypes. In a previous study, frequent intragenic variation of the gB gene was shown. The aim of this study was to analyse whether gB variation was due to homologous recombination. The gB gene of DNA extracts derived from the peripheral blood leukocytes of 14 immunosuppressed patients was amplified by PCR and cloned. Three variable sites of gB were analysed by restriction fragment analysis and DNA sequencing and compared with published prototypic strains. In three patients doubly infected with two distinct HCMV gB strains, prototypic (60-85%) and non-prototypic recombinant strains (5-40%) were detected. To demonstrate that homologous recombination is driving HCMV gB variability, cells were coinfected with plaque-purified prototypic gB strains and recombinant gB genes were selectively amplified by PCR. gB recombinants were detected after 15 days of coculture and cross-over sites were determined by sequencing. These data indicate that homologous recombination contributes to the variability of the gB gene in vitro and in vivo.

Intragenic variability of human cytomegalovirus glycoprotein B in clinical strains.

Human cytomegalovirus (HCMV) strains can be classified into four glycoprotein B (gB) genotypes, and there has been evidence of differences in viral virulence. In this study, intragenic variability of HCMV gB strains was analyzed. The gB gene was amplified by nested polymerase chain reaction using samples from immunosuppressed patients. The genotype of fragments corresponding to the cleavage site of gB was determined by restriction fragment analysis; fragments corresponding to the N- and C-termini (gBn and gBc) were sequenced and compared with published sequences. At the cleavage site, the four known genotypes were found. Typing revealed four major genotypes at the N-terminus and two at the C-terminus. In 22 of 44 strains, the gB type determined at the cleavage site was different from the gBn or gBc type (or either), indicating that intragenic variability within the gB gene occurs frequently.

Glycoprotein B genotype correlates with cell tropism in vivo of human cytomegalovirus infection.

Human cytomegalovirus (HCMV) strains can be classified into four genotypes of the glycoprotein B (gB). In a previous study, the gB genotype 1 was found more frequently in bone marrow transplant recipients with nonfatal HCMV infection than in patients who died from HCMV disease [Fries et al. (1994): Journal of Infectious Diseases 169:769-774]. The distribution and cell tropism of different gB types in vivo were investigated. The gB type of HCMV was determined in blood or urine specimen from 76 organ and 47 bone marrow transplant recipients using PCR and restriction fragment length polymorphism (RFLP). The leukocyte populations (polymorphonuclear leukocytes, monocytes, T lymphocytes, non-T lymphocytes) of 20 viremic patients were purified by a fluorescence-activated cell sorter (FACS) and examined for HCMV infection by PCR. Sequence analysis of four randomly selected strains showed that gB types were similar to published sequences and no atypical gB types were found. Within the compartments blood and urine, the gB types were almost equally distributed, whereas the gB type 1, in contrast to gB types 2 and 3, did not infect T lymphocytes in vivo. These data show that the gB type correlates with viral tropism in vivo and thus provides further evidence that the gB variation may indeed influence the virulence of HCMV.

Lipoprotein trafficking in vascular cells. Molecular Trojan horses and cellular saboteurs.

During the pathogenesis of atherosclerosis, inflammatory cells such as the monocyte-derived macrophage accumulate in the vessel wall where they release cytokines. Initially, cytokines may assist in CE removal of lipoprotein-derived cholesterol/CE hydrolysis to clear intracellular lipid. When plasma levels of LDL become elevated, the vessel wall becomes lipid-engorged over time because it is unable to traffick the large amounts of endocytosed LDL-CE from the cell. In addition, lipoprotein entrapment by the extracellular matrix can lead to the progressive oxidation of LDL because of the action of lipoxygenases, reactive oxygen species, peroxynitrite, and/or myeloperoxidase. A range of oxidized LDL species is thus generated, ultimately resulting in their delivery to vascular cells through several families of scavenger receptors (Fig 1). These molecular Trojan horses and cellular saboteurs once formed or deposited in the cell can contribute to, and participate in, formation of macrophage- and smooth muscle-derived foam cells. A lipid-enriched fatty streak along the vessel wall can ensue. In addition to foam cell development, products of LDL peroxidation may activate endothelial cells, increase smooth muscle mitogenesis, or induce apoptosis because of the effects of oxysterols and products of lipid peroxidation (Fig 1). Because antioxidant defenses may be limited in the microenvironment of the cell or within LDL, the oxidation process continues to progress. Enzymes associated with HDL such as PAF acetylhydrolase and paraoxonase can participate in the elimination of biologically active lipids, but diminished cellular antioxidant activity coupled with low levels of HDL may allow acceleration of the clinical course of vascular disease. There is still much to be learned about how modified LDL initiate cellular signals that lead to inflammation, mitosis, or cholesterol accumulation. The present challenges include elucidation of the key signaling events that regulate lipoprotein-derived cholesterol trafficking in the vessel wall, which can impact on the pathogenesis of vascular disease.

Amino terminus of apolipoprotein B suffices to produce recognition of malondialdehyde-modified low density lipoprotein by the scavenger receptor of human monocyte-macrophages.

Malondialdehyde, a product of lipid peroxidation, produces threshold conversion of low density lipoprotein (LDL) to a form recognized by type I and type II scavenger receptors of monocyte-macrophages. To investigate whether localized domains of human apoB-100 protein provide recognition determinants, we tested the ability of several different apoB-bearing particles to interact with the scavenger receptor of human monocyte-macrophages. Genetically engineered, carboxyl-terminally truncated apoB proteins assembled into lipoprotein form were labeled by fluorescent dye. Fluorescence microscopy and quantitative fluorescent spectrophotometry showed that purified particles containing as little as 23% of the apoB amino-terminus were internalized by the scavenger receptor after, but not before, malondialdehyde modification. There was no recognition of the particles by the LDL receptor. Similar results were obtained with human plasma LDL homozygous for carboxyl-terminally truncated apoB-45.2. Liposome-incorporated fusion protein containing apoB residues 547-735 displayed specific uptake by the scavenger receptor without modification by malondialdehyde. In contrast, fusion proteins containing apoB residues 3,029-3,133 or a short amino terminal segment failed to interact. Thus, primary sequence presented by residues 1-1,084 sufficed to produce recognition of modified LDL by the scavenger receptor. These receptor-combining domains were sequestered when secreted in lipoprotein form and were expressed upon malondialdehyde modification. When packaged exogenously in liposome form, fusion protein containing apoB residues 547-735, containing approximately 4% of the primary sequence, mediated scavenger receptor-dependent uptake and hydrolysis. Our findings provide an additional function or the amino-terminal region of apoB and demonstrate that primary sequence presented by the first 2% of apoB-100 protein suffices to produce recognition on malondialdehyde-modified LDL by the scavenger receptor of human monocyte-macrophages.

A cytotoxic electronegative LDL subfraction is present in human plasma.

By using fast protein liquid chromatography, we isolated from human plasma a minor electronegative LDL subfraction designated LDL(-). After immunoaffinity chromatography against apolipoprotein (apo)(a) and apo A-I, LDL(-) represented 6.7 +/- 0.9% (mean +/- SD; n = 18) of total LDL. Compared with the major LDL subfraction, designated LDL(+), LDL(-) contained similar amounts of thiobarbituric acid-reactive substances, conjugated dienes, and vitamin E and had a similar lipid/protein ratio and mean density. Moreover, the apo B of LDL(-) was not aggregated and its LDL receptor-binding activity was slightly increased. These results were consistent with the nonoxidized nature of LDL(-). LDL(-) showed increased contents of sialic acid (38.1 +/- 5.2 versus 28.9 +/- 3.3 nmol/mg protein; n = 7; P < .01), apo C-III (1.43 +/- 0.21% versus 0.14 +/- 0.04%; n = 7; P < .01), and apo E (1.64 +/- 0.26% versus 0.10 +/- 0.05%; n = 7; P < .0005). Compared with LDL(+), LDL(-) displayed enhanced cytotoxic effects on cultured human umbilical vein endothelial cells, as shown by lactate dehydrogenase assay (P < .003; n = 6), neutral red uptake (P < .02; n = 6), and morphological studies. We also studied the relationship of LDL(-) to age and plasma lipid levels in 133 subjects. The percentage of contribution of LDL(-) to total plasma LDL correlated with age (P < .05), total cholesterol (P < .05), and LDL cholesterol (P < .003). In conclusion, this study shows that LDL(-), a circulating human plasma LDL, is an electronegative native LDL subfraction with cytotoxic effects on endothelial cells. This subfraction, which correlates positively with common atherosclerotic risk factors, might induce atherogenesis by actively contributing to alteration of the vascular endothelium.

Partial characterization of leukocyte binding molecules on endothelial cells induced by minimally oxidized LDL.

Treatment of rabbit aortic endothelial cells, human umbilical vein endothelial cells, and human aortic endothelial cells for 4 hours with minimally oxidized low-density lipoprotein (MM-LDL) induced the adhesion of monocytes but not neutrophils or lymphocytes to these cells. This induction was blocked by inhibitors of glycoprotein synthesis (cycloheximide and tunicamycin), and binding was abolished by treatment of cells with low levels of trypsin, suggesting that the binding molecule(s) is a protein. There was no increase in binding of antibodies to E-selectin, vascular cell adhesion molecule-1 (VCAM-1), or intercellular adhesion molecule-1 (ICAM-1) after treatment of cells with MM-LDL. Treatment of endothelial cells with Fab fragments of antibody to monocyte chemotactic protein-1 or to fibronectin did not block monocyte binding. Several sugars (lactose-1-phosphate, maltose-1-phosphate, and N-acetylglucosamine) inhibited monocyte binding to cells treated with MM-LDL, but binding was not blocked by mannose-6-phosphate, fructose-6-phosphate, glucose-1-phosphate, or glucose-6-phosphate. EDTA or EGTA treatment inhibited binding, which was restored by adding either calcium or magnesium. We conclude that the binding of monocytes to endothelial cells induced by a 4-hour treatment with MM-LDL is caused by a binding molecule(s) other than E-selectin, VCAM-1, or ICAM-1 and that carbohydrate chains on the monocytes or the endothelium play a role in binding.

Lipoprotein (a) displays increased accumulation compared with low-density lipoprotein in the murine arterial wall.

Lipoprotein (a) (Lp(a)) is known to be an independent risk factor for cardiovascular disease, but the mechanisms by which it contributes to this disease remain unclear. Current evidence indicates that the closely related plasma particle, low-density lipoprotein (LDL), may initiate atherosclerosis through deposition in the arterial wall. This study has compared the ability of both lipoproteins to enter and accumulate within the arterial wall. Experiments were conducted in vivo with animals from two strains of mice: C57BL/6 mice, which develop fatty streak lesions upon challenge by a high-fat diet, and C3H/HeJ mice, which are resistant to lesion formation. Animals from both strains were maintained up to 16 weeks either on chow or high-fat diet. The mice were intravenously injected with 125I-labeled human Lp(a) or 125I-labeled human LDL in equimolar amounts and the lipoprotein allowed to circulate in vivo for 2 or 24 h. Transverse sections of the aortic root including sites of predilection for lesion formation at the commissures of the valve were prepared and examined after autoradiography. The autoradiographic grains over lesions and histologically uninvolved areas were enumerated and compared after normalization. Both Lp(a) and LDL demonstrated nearly ten times greater accumulation in lesions compared with histologically uninvolved areas from C57BL/6 mice. Analyses of histologically uninvolved areas from both strains of mice showed a significantly higher accumulation of Lp(a) than LDL. Finally, significantly higher accumulations of both Lp(a) and LDL occurred in the histologically uninvolved intima and subintima of lesion-prone C57BL/6 mice as compared with lesion-resistant C3H/HeJ mice after 5 weeks on the diets. We propose that enhanced accumulation of Lp(a) in the arterial wall accounts, in part, for the increased risk of cardiovascular disease.

Malondialdehyde modification of lipoprotein(a) produces avid uptake by human monocyte-macrophages.

Increased plasma levels of the apoB-100-containing lipoprotein(a) (Lp(a)) are associated with an increased risk for atherosclerosis and myocardial infarction, but the mechanisms by which lipoprotein(a) may accelerate these processes remain obscure. In this study we have investigated the impact of the association of apoprotein(a) with the low density lipoprotein (LDL)-like Lp(a) particle upon specificity of receptor recognition after lipoprotein modification by malondialdehyde or transition metal-induced oxidation. We have determined that radioiodination labels both apoprotein components of Lp(a), that malondialdehyde modification produces an anionic lipoprotein comparable to native Lp(a) in Stokes' radius, and that N,N'-disubstituted 1-amino-3-iminopropene derivatives preferentially cross-link apoprotein(a) to apoB-100 protein. Like LDL, native Lp(a) is recognized in human monocyte-macrophages by the LDL receptor. Like LDL, progressive modification of Lp(a) by malondialdehyde abolishes lipoprotein recognition by the LDL receptor and produces uptake and hydrolysis by the scavenger receptor of human monocyte-macrophages. We propose that intimal retention of Lp(a) by extracellular components of the atherosclerotic reaction places the lipoprotein in a microenvironment favoring subsequent peroxidative modification. The chronic production of lipid peroxide-modified Lp(a) together with unmitigated cellular clearance by scavenger receptors may contribute to the accumulation of lipoprotein-derived lipid in macrophage-derived foam cells of the atherosclerotic reaction.

Identification of the cardiac sarcolemmal Na(+)-Ca2+ exchanger using monoclonal antibodies.

We have previously partially purified the sarcolemmal Na(+)-Ca2+ exchange protein and produced rabbit polyclonal antibodies to the exchanger (Philipson, K.D., Longoni, S., Ward, R. 1988. Biochim. Biophys. Acta 945:298-306). We now describe the generation of three stable murine hybridoma lines which secrete monoclonal antibodies (MAb's) to the exchanger. These MAb's immunoprecipitate 50-75% of solubilized Na(+)-Ca2+ exchange activity. The MAb's appear to be reactive with native conformation-dependent epitopes on the Na(+)-Ca2+ exchanger since they do not react on immunoblots. An indirect method was used to identify Na(+)-Ca2+ exchange proteins. A column containing Na(+)-Ca2+ exchanger immobilized by MAb's was used to affinity purify the rabbit polyclonal antibody. The affinity-purified polyclonal antibody reacted with proteins of apparent molecular weights of 70, 120, and 160 kDa on immunoblots of sarcolemma. The data provide strong support for our previous association of Na(+)-Ca2+ exchange with these proteins.