Embryonic stress hypothesis of teratogenesis

Hyperthermia: malformations to chaperones

Hyperthermia has been known to induce malformations in numerous animal models as well being associated with human abnormalities. This was apparent particularly when the hyperthermia exposure was during the early stages of neural development. Although it was recognized relatively early that these exposures induced cell death, the specific molecular mechanism of how a brief heat exposure was translated in to specific cellular functions remains largely unknown. While our understanding of the events that govern how cells react to heat, or stresses in general, has increased, there is much that remains undiscovered. In this brief review, animal and clinical observations are outlined as are some of the scientific explorations that were undertaken to characterize, define, and better understand the morphological, biochemical, and molecular effects of hyperthermia on the developing embryo.

IDENTIFICATION AND VALIDATION OF DIFFERENTIALLY EXPRESSED GENES IN NEURAL TUBE DEFECTS OF GOLDEN HAMSTER INDUCED BY HYPERTHERMIA USING SUPPRESSION SUBTRACTIVE HYBRIDIZATION

Hyperthermia during early pregnancy can induce neural tube defects (NTD) in embryos. In order to identify the differentially expressed genes that participate in this pathologic course, the authors performed suppression subtractive hybridization in two directions: forward and reverse. Neural tube tissues from golden hamster of normal and hyperthermia groups are used as the samples. As a result, several down-regulated genes were revealed and according to the function of their protein products they were classified into four categories: ribosomal proteins, metabolic enzymes, transcription and translation related factors, and others. On the other hand, the study found that two up-regulated gene fractions were of the same sequence and homology analysis shows that they are homologous to phosphoglycerate kinase 1 (pgk1). Of all these genes, differential expression patterns were confirmed by Northern blot analysis. The study results show that the genes identified have different expression and are stongly related to NTD induced by hyperthermia.

Birth defects and anti-heat shock protein 70 antibodies in early pregnancy

Abstract It has been suggested that induction of the heat shock response in the mammalian embryo during the critical period of organogenesis can result in anatomical malformation. We measured serum heat shock protein 70 (Hsp70), anti-Hsp70, and anti-Hsp60 in samples taken from expectant mothers at 16 weeks gestation. Samples from women whose babies were born with a birth defect (n = 30) were compared with controls who gave birth to healthy babies (n = 46). Anti-Hsp70 levels were significantly elevated in patients who later gave birth to babies with cleft lip or palate or neurological abnormalities (n = 10): 260 (223-406) microg/mL compared to 150 (88-207) microg/mL in controls (P < 0.001). No significant differences were found in serum Hsp70 and anti-Hsp60 levels between cases and controls. This finding of increased maternal anti-Hsp70 in patients who later gave birth to babies with these abnormalities suggests a previous stressful event may have contributed to the pathogenesis. Further work is required to determine whether Hsp70 has a direct or indirect role in this pathogenesis or whether anti-Hsp70 is simply a marker of a prior increase in Hsp70 due to a physiological stress that itself resulted in the damage. This work is consistent with previous studies showing a buffering role for Hsps in evolution.

Heat stress induced histopathology and pathophysiology of the central nervous system

The number of reports on the effects of heat stress is still increasing on account of the temperature is one of the most encountered stressful factors on the different biological systems. Because the heat stress (HS) considered a model of thermal injury to the central nervous system (CNS), the purpose of this review was to assess the histopathological changes of HS on CNS. Also, this review emphasized that the heat stress may retard partially the degree of the postnatal neurogenesis and growth of CNS. Taken together, owing to one of the most important functions of heat shock protein is to protect the organisms from the deleterious effects of temperature, thus, it can be hypothesized that the formation of heat shock proteins may be related to the deleterious effect of HS. On the other hands, the alterations of neurotransmitters in the central nervous system might be involved in the physiological and biochemical responses that occur during heat stress. The hypothalamic monoaminergic systems play an important role in the thermoregulation through regulate the heat production and heat dissipation. In addition, the disturbance in the biochemical variables due to the high temperature may be the cause of the histopathological changes and the partial retardation in CNS and the reverse is true. Thus, further studies need to be done to emphasize this concept.

Genetic dissection of hyperthermia-induced neural tube defects in mice

BACKGROUND

Maternal hyperthermia has been shown to induce neural tube defects (NTD) in humans and in experimental animal systems. We report the first genetic dissection of maternal hyperthermia-induced NTD in mice.

METHODS

After maternal exposure on E8.5 to 43 degrees C water bath for 10 min, we observed exencephaly frequencies among E15.5-17.5 fetuses from the following crosses and backcrosses, SWV/Fnn(SWV)xSWV, C57BL/6J(C57)xC57, SWVxC57 (F1), F1xSWV and SWVxF1.

RESULTS

The fetuses with maternal hyperthermia exposure developed exencephaly in a strain-dependent manner and the exencephaly frequencies among the above crosses were 46.2, 14.3, 13.6, 11.3, and 27.0%, respectively, expressed over total live fetuses. The fetal death rates were 47.3, 24.6, 37.1, 4.3, and 35.5%, respectively, expressed over total implants.

CONCLUSION

The data demonstrate that a single fetal genetic locus, plus a maternal effect, have likely caused the strain differences in the susceptibility to hyperthermia-induced exencephaly. A maternal effect alone may have caused the higher prenatal mortality rates in the SWVxF1 cross versus the reciprocal cross. Analysis of gender ratios among those affected from these crosses excludes an X- or Y-linked effect in causing the higher numbers of affected females.

Identification of antibodies to heat shock proteins 90 kDa and 70 kDa in patients with schizophrenia

OBJECTIVES

Recent reports of antibodies to heat shock proteins 60kDa (HSP60) and HSP70 suggested that antibodies to the heat shock protein that plays a protective role against environmental stresses in a cell might be related to the pathogenesis of schizophrenia, although the antibody to HSP90 had not yet been identified in patients with schizophrenia. In this study, we tried to elucidate the specific involvement of the autoimmunity to HSPs in the pathogenesis and development of schizophrenia.

METHODS

Antibodies to HSP90 and HSP70 in 90 patients with schizophrenia and in 83 normal controls were measured by enzyme-linked immunosorbent assay (ELISA) coupled with the avidin-biotin system. In the patients, the association between antibody levels and clinical variables were sought. In addition, changes in antibody levels after treatment with antipsychotic medication were investigated.

RESULTS

Eighteen (20.0%) of the 90 patients showed 'high' levels of antibody to HSP90 above a cutoff value, and 28 (31.1%) of those showed 'high' antibody levels to HSP70. On the other hand, only four (4.8%) of the normal controls showed 'high' HSP90 antibody levels, and one (1.2%) of these showed 'high' antibody level to HSP70. The distribution of elevated HSP90 antibody was significantly associated with that of elevated HSP70 antibody in the patients with schizophrenia. The patients with 'high' levels of antibody to HSP70 showed higher initial Brief Psychiatric Rating Scale (BPRS) scores and showed greater clinical improvement than those with 'low' levels, while the patients with 'high' levels of antibody to HSP90 did not. The frequency of patients with high levels of antibody to HSP70 was decreased significantly after 6 weeks of antipsychotic treatment, while the frequency of patients with high levels of antibody to HSP90 was not.

CONCLUSIONS

Our results presented the presence of abnormal immune reactivity involving antibody to HSP90 and antibody to HSP70 in a subset of patients with schizophrenia. Differential patterns of distribution, of the association with clinical symptom severity, and of the changes of levels with treatment suggested the possibility that these two antibodies might be involved specifically in the pathogenesis of schizophrenia.

Teratogen update: gestational effects of maternal hyperthermia due to febrile illnesses and resultant patterns of defects in humans

This review has covered the pertinent literature concerning the teratogenic effects of hyperthermia in man and experimental animals. This is the first teratogen that was initially discovered in animals and then subsequently found to be a cause for concern in humans when similar patterns of defects were observed. Hyperthermia is a physical agent with a dose-response curve for abortions and malformations, but these effects can be mitigated in some circumstances by the heat shock response (HSR). We have reviewed the known functions of HSR and provided some insight into why embryos have some protection following an initial dose of heat, if it is sufficient to initiate the response. Thus, by reviewing the effects of hyperthermia in experimental animals, as well as malformative and protective mechanisms of teratogenesis, we have attempted to understand the effects of human hyperthermia teratogenesis.

Developmental and reproductive toxicity of inorganic arsenic: animal studies and human concerns

Information on the reproductive and developmental toxicity of inorganic arsenic is available primarily from studies in animals using arsenite and arsenate salts and arsenic trioxide. Inorganic arsenic has been extensively studied as a teratogen in animals. Data from animal studies demonstrate that arsenic can produce developmental toxicity, including malformation, death, and growth retardation, in four species (hamsters, mice, rats, rabbits). A characteristic pattern of malformations is produced, and the developmental toxicity effects are dependent on dose, route, and the day of gestation when exposure occurs. Studies with gavage and diet administration indicate that death and growth retardation are produced by oral arsenic exposure. Arsenic is readily transferred to the fetus and produces developmental toxicity in embryo culture. Animal studies have not identified an effect of arsenic on fertility in males or females. When females were dosed chronically for periods that included pregnancy, the primary effect of arsenic on reproduction was a dose-dependent increase in conceptus mortality and in postnatal growth retardation. Human data are limited to a few studies of populations exposed to arsenic from drinking water or from working at or living near smelters. Associations with spontaneous abortion and stillbirth have been reported in more than one of these studies, but interpretation of these studies is complicated because study populations were exposed to multiple chemicals. Thus, animal studies suggest that environmental arsenic exposures are primarily a risk to the developing fetus. In order to understand the implications for humans, attention must be given to comparative pharmacokinetics and metabolism, likely exposure scenarios, possible mechanisms of action, and the potential role of arsenic as an essential nutrient.

Molecular/cellular biology of the heat stress response and its role in agent-induced teratogenesis

Available data indicate that heat shock proteins act as chaperones under non-stress conditions by assisting in: (1) the folding of newly synthesized proteins, (2) the intracellular translocation of proteins, and (3) the function of other proteins. As we gain additional information concerning cellular physiology, we may find that heat shock proteins play a key role in many additional cellular functions. When cells experience thermal or chemical stress, heat shock proteins take on a new role, conserved from bacteria to humans, of protecting cells from the detrimental effects of stress. This latter role takes on added significance for the embryo in which the developmental program must be read linearly, with little opportunity to cycle backward to complete a missed segment of the program. Although circumstantial evidence clearly implicates heat shock proteins in protecting embryos from thermal stress, definitive evidence is still lacking. The challenge for the future is to obtain such definitive data. Ideally, such information will lead to new therapeutic paradigms that will afford protection to the human embryo/fetus exposed to thermal/chemical stress.

Heat stress lipids and schizophrenia

The neurodevelopmental hypothesis of schizophrenia implicates abnormal or disrupted neural growth during embryogenesis. It is postulated here that stress-inducing agents acting upon a compromised cellular system resulting from abnormal plasma membrane lipids could effect the neuronal abnormalities observed in schizophrenia. The heat stress response is induced by exposure to hyperthermia as well as a variety of other agents. The response to these agents includes the cessation of most transcriptional and translational activities, accompanied by the induction of a highly specific set of proteins. A concomitant reduction in metabolic activity including cell cycle delays is also observed. Much of the enormous literature on the heat stress response concentrates on protein and DNA interactions, especially with regard to transcriptional control. However, a variety of lipids are intrinsically involved in the heat stress response. This paper will provide a brief introduction to the heat shock proteins and will explore the roles that lipids play in the heat shock response.

Screening for reproductive toxicity in Fundulus heteroclitus by genetic expression profiling

Abstract Potentially teratogenic agents enter the environment at a rate that greatly exceeds current capabilities to effectively evaluate their reproductive toxicities. This is due, in part, to costly, labour-intensive methodologies involving mammalian embryonic screening assays that are currently in use worldwide. Therefore, we sought to develop a rapid, less expensive screening system with which to identify molecular biomarkers of teratogenicity using a non-mammalian system. Embryos of the topminnow, Fundulus heteroclitus, offer several advantages in terms of reproductive toxicity screening efficiency as compared with mammalian embryonic systems. These embryos are easily manipulated and develop normally at ambient temperature in air, water, or air-saturated mineral oils, making them readily adapted for field studies. In the present study, developing f. heteroclhs embryos were exposed to teratogenic concentrations of sodium valproate (VPA) or arsenic acid (arsenate), and the frequency and types of induced malformations were evaluated. Using in situ transcription and antisense RNA (aRNA) amplification procedures (IST/aRNA), we attempted to correlate the teratogenic outcomes to specific alterations in the expression of a panel of developmentally regulated genes. Preliminary studies identified treatment concentrations of arsenate and VPA that induced abnormal development in 95 % of the surviving embryos. Among the F. heteroclitus embryos, the structural defects most commonly induced by these compounds were cardiac and neural tube malformations. The genetic expression profiles revealed a number of genes whose expression levels were significantly altered by exposure to the test compounds. Molecular analysis of f. heteroclitus embryonic development represents a novel, inexpensive approach to screen for potential teratogens, and identify genes whose expression patterns may be used as biomarkers, or indicators, of teratogenicity.

Retinoic acid-induced stress protein synthesis in the mouse

We have previously demonstrated that stress proteins (SPs) are synthesized in tissues in which malformations are later observed following treatment with the developmental toxicant, retinoic acid (RA), on day 11 of gestation (GD 11). These proteins were not synthesized in tissues which did not present with malformations near partuition. The purpose of the present investigation was to determine if this correlation between early SP synthesis and later malformation was present at other times during gestation. CD-1 strain mice were dosed orally with corn oil or 100 mg/kg body weight RA on GD 10 or 13. Some of the mice in each group were given an intraperitoneal injection of 3H-leucine to label embryonic protein synthesis one hour after dosing with RA. These animals were sacrificed 1.5 hour later, and embryonic protein synthesis was determined by two-dimensional gel electrophoresis followed by autoradiography. Other animals in each group were sacrificed on day 17 of gestation, and fetuses were examined for the presence of malformations. Following treatment with RA on day 10 of gestation, malformations were observed in the forelimbs, the hindlimbs and the tail; heart defects were not observed. SPs of 20-25,000 and 90,000 relative molecular mass (Mr) were synthesized in the forelimb bud and tail; in addition, a second low molecular weight (20-25,000) and a 84,000 Mr SPs were synthesized in forelimb buds. No SPs were synthesized in the hindlimb bud or the heart. Following RA treatment on GD 13, cleft palate was observed in 58% of fetuses; no other malformations were found. Proteins of 34,000, 84,000 and 90,000 Mr were synthesized in craniofacial tissue; SPs were not observed in forelimb bud, hindlimb bud, heart or tail tissues at this time. Therefore, it appears that there may be a correlation between tissue-specific SP synthesis early in organogenesis and the presence of a malformation later in gestation.

Maternal restraint stress-enhanced teratogenicity of all-trans-retinoic acid in CD-1 mice

The present study combined maternal restraint stress with a teratogenic agent, all-trans-retinoic acid (tRA). Five treatment groups were used initially: (1) vehicle (corn oil) control [C], (2) food/water-deprived [FWD], (3) tRA only [tRA], (4) restraint only [R], and (5) tRA plus restraint [tRA+R]. Mated CD-1 mice in groups 3 and 5 were given 20 mg/kg tRA po. Mice in groups 4 and 5 were restrained in the supine position for 12 hr (9:00 a.m. to 9:00 p.m.), and the FWD group mice were deprived during the same time period. The tRA+R mice were dosed immediately prior to the 12-hr restraint period. All treatments were administered on gestation day (GD) 9 (copulation plug = day 1). On GD 18, all females were killed and subjected to teratological examination. The incidences of resorptions, short tails, bent tails, fused ribs, and fused vertebrae were significantly increased in the tRA+R group, in comparison with all other groups. Spina bifida was observed only in the tRA+R group. The current results, combined with those of earlier studies with other agents, support the likelihood that maternal stress can exacerbate adverse effects of chemical teratogens on mouse development.

The influence of earthquake-induced stress on human facial clefting and its simulation in mice

A large earthquake (8-9 on the Richter scale) and a series of aftershocks took place on 2 March 1985 in Santiago, Chile. The characteristics of over 22,000 births registered in three public hospitals in the same year were reviewed. A significant increase in the rate of facial clefts was found; 2.01 per 1000 births in contrast to 1.6 per 1000 births in previous years. The increase was greater in those born in September: 3.8 per 1000 births. This increase in clefting could be related to the effects of stress in mothers induced by the earthquake, and to test this hypothesis 13.5-day-old embryos from two inbred mouse strains, A/Sn and C57BL/10, were subjected to a similar stress using a vibrator cage to imitate the main shock and the first five replicas of the earthquake. The same intensity and duration of shock as in the original earthquake were applied. The results were 19.8% cleft palates in stressed A/Sn mice and no clefting in C57BL/10. This was highly significant in A/Sn mice (chi 2 = 19.9; P < 0.001) but not in C57BL/10. No clefting was found in controls in both strains. A surprising finding was the proportion of resorbed embryos in the stressed groups, which increased from 8.3 to 49.3% in A/Sn and from 5.8 to 48.3% in C57BL/10. It is known that A/Sn mice are genetically sensitive to cleft palate induction by cortisone, while C57BL/10 are not. These findings in mice support the stress hypothesis for the increase in cleft palate observed in humans. The increase in resorbed embryos in both strains also suggests an effect on stress.

Induction of a heat shock response (HSP 72) in rat embryos exposed to selected chemical teratogens

A monoclonal antibody to the 72 kD heat shock protein (HSP 72), Western blot analysis and 2-D gel electrophoresis/autoradiography were used to determine whether selected chemical teratogens induced the synthesis and accumulation of HSP 72 in postimplantation rat embryos exposed in vitro. The chemical teratogens studied include N-Acetoxy-2-acetylaminofluorene (N-Ac-AAF), cadmium chloride (CAD), cyclophosphamide (CP), sodium arsenite (AS), and sodium salicylate (SAL). Exposures to test chemicals were selected that produced obvious embryotoxicity characterized by abnormal development and growth retardation. Of the five chemical teratogens studied, AS and SAL induced the synthesis and accumulation of HSP 72 in day 10 rat embryos. The kinetics of HSP 72 accumulation, however, differed between AS- and SAL-treated embryos. Maximal levels of HSP 72 were observed 24 hours after AS exposure and 10 hours after SAL exposure. N-Ac-AAF, CD, and CP induced obvious embryotoxicity; however, none of these chemical teratogens induced HSP 72 at any of the timepoints assayed. Although only a small sample of chemical teratogens was studied, our results suggest that the heat shock response, characterized by the synthesis and accumulation of HSP 72, is not a general biomarker for chemical teratogens.

In utero exposure to caffeine causes delayed neural tube closure in rat embryos

We have investigated the effect of caffeine on embryo growth and development. Caffeine (25 mg/kg) was administered on gestation day (g.d.) 8-9 and the embryos examined histologically 24 h after the final dose. The crown-rump length of caffeine treated embryos (1.92 +/- 0.08 mm) was significantly smaller (P < 0.001) than the controls (2.91 +/- 0.26 mm) as was the circumferential length (caffeine vs. controls, 3.79 +/- 0.16 mm vs. 6.03 +/- 0.61 mm; P < 0.001). Additional measures, such as development of the heart, eye, and limb buds, were also reduced in the caffeine treated embryos. The most striking difference between the control and caffeine treated embryos was the larger proportion of treated embryos with regions of open neural tube. This was most marked in the caudal region of the embryos where 91% of treated embryos had regions of open neural tube compared with 14% of controls. The amount of open neural tube in any individual caffeine treated embryo did not relate to the crown-rump or circumferential length of that embryo nor was the effect restricted to particular litters. These results indicate that caffeine had a significant effects on embryonic growth and development.

Stress protein synthesis induced by cadmium-cysteine in rat kidney

Biomarkers are important tools which enable toxicologists to reliably predict and detect exposures to xenobiotics and resultant cell injury, ultimately improving risk assessments. Since the de novo synthesis of stress proteins can be detected early after exposure to some agents, analysis of toxicant-induced changes in gene expression, i.e. alterations in patterns of protein synthesis, may be useful to develop as biomarkers of exposure and toxicity. We are utilizing various xenobiotics as tools to study stress protein synthesis in target organs in order to evaluate the target tissue-specificity of this response. Previous data from this laboratory have demonstrated that induction of stress proteins in rat liver, but not kidney, after acute exposure to CdCl2 precedes hepatoxicity. Since kidney is a target tissue after chronic Cd exposure, it was of interest to examine stress protein synthesis in this tissue. However, dose-limiting hepatotoxicity precluded this evaluation. Cd complexed with molecules such as cysteine (cys) or metallothionein has been used in acute dosing regimens as a tool in order to study the nephrotoxicity of Cd. Therefore, this study was undertaken in order to evaluate Cd-induced stress protein synthesis in an important tissue known to be injured after chronic exposure, i.e. kidney. Specific objectives included comparing stress protein synthesis in rat kidney and liver after acute exposure to Cd-cys and CdCl2, determining the Cd threshold concentration for renal stress protein synthesis and assessing the relationship between stress protein synthesis and nephropathy. Male rats were exposed to equivalent doses of Cd as CdCl2 or Cd-cysteine (molar ratio Cd:cys = 1:15). Kidney Cd concentrations increased 5-fold after i.v. injection of Cd-cys compared to CdCl2, mimicking Cd distribution following chronic exposure. After exposure to Cd, tissue slices were incubated with 35S-methionine. Slices were subsequently homogenized and centrifuged, and the 16,000 g supernatants were subjected to SDS-polyacrylamide gel electrophoresis. Proteins which had incorporated 35S-methionine were detected by autoradiography. De novo synthesis of 70, 90 and 110 kDa proteins was enhanced in liver, but not in kidney, 4 h after injection of 2 mg Cd/kg as CdCl2. In contrast, dose-related increases in synthesis of these proteins were observed in kidney 4 h after injection of 1 and 2mg Cd/kg as Cd-cys, but not at lower dosages. In addition, synthesis of a 68 kDa kidney protein was inhibited at 2 mg Cd/kg as Cd-cys. The threshold for Cd-induced stress protein synthesis was shown to be between 4 and 8 micrograms Cd/g tissue.(ABSTRACT TRUNCATED AT 400 WORDS)

A comparison of sodium arsenite- and hyperthermia-induced stress responses and abnormal development in cultured postimplantation rat embryos

Acute exposures to sodium arsenite (50 microM) were embryotoxic in day 10 rat embryos exposed in vitro. Sodium arsenite-induced embryotoxicity was characterized by decreased growth (crown-rump length, somite number, and embryo protein content) and abnormal development (hypoplastic prosencephalon, abnormal somites, and abnormal flexion of the tail). At embryotoxic exposures, sodium arsenite also induced the synthesis of three heat shock proteins (hsps), one of which is recognized by a monoclonal antibody specific for the heat-inducible hsp 72. In addition, sodium arsenite induced the accumulation of heat-inducible hsp 70 mRNA. Although the abnormal morphologies induced by sodium arsenite and hyperthermia appear to be different, the stress response as measured by the synthesis of hsps, the accumulation of hsp 72 protein, and the accumulation of hsp 70 mRNA is similar in embryos exposed to these two embryotoxic agents. Thus, sodium arsenite and hyperthermia both induce a stress response; however, the relationship between the induction of a stress response and the subsequent abnormal development that ensues is unclear.

Rapid in vivo reactivation of herpes simplex virus in latently infected murine ganglionic neurons after transient hyperthermia

A rapid and physiologically relevant hyperthermia-based induction procedure has been utilized to develop an in vivo model of induced herpes simplex virus (HSV) reactivation in outbred Swiss Webster mice. This procedure was found to efficiently reactivate latent virus from both trigeminal and lumbosacral ganglia. Examination of the time between hyperthermia and virus production demonstrated that detectable levels of infectious virus were present in ganglia as soon as 14 h posttreatment, with peak percent recoveries at 24 h. These data indicated that the switch from latent to active viral gene transcription occurred rapidly following treatment. Immunohistochemical staining for HSV type 1 antigens revealed rare antigen-positive ganglionic neurons 24 h postinduction. HSV antigens were not detected in any other cell type, and lateral spread of the infection was not observed. This is the first report of the detection of HSV antigens in vivo following induced reactivation in the intact nervous system and demonstrates that the neuron is the site of infectious virus production. In addition, our data strongly suggest that at least some neurons in which HSV antigens are detected during reactivation do not survive. Because the temporal and spatial characteristics of HSV reactivation have been clearly defined, this model is uniquely suited for the molecular dissection of the reactivation process.

Relationship between stress protein induction in rat kidney by mercuric chloride and nephrotoxicity

Adverse environmental stimuli increase the synthesis of a class of proteins referred to as stress proteins. The effect of mercuric chloride, a model nephrotoxin, on protein synthesis in male rat kidney has been evaluated. Renal slices from exposed rats were incubated with [35S]methionine for 1 hr and subjected to SDS-PAGE, after which 35S-labeled proteins were detected by autoradiography. Enhanced de novo synthesis of 70- and 90-kDa relative molecular mass (M(r)) proteins were detected 2 hr after exposure to 1 mg Hg/kg, with maximum activity occurring at 4-8 hr. By 16 hr postinjection, synthesis of these two proteins had decreased. Dose-related increases in synthesis of these proteins, and of a 110-kDa protein, were observed 4 hr after i.v. injection of 0.25, 0.5, and 1.0 mg Hg/kg, with concomitant inhibition of synthesis of proteins of M(r) 38 and 68 kDa. At a dose of 1 mg/kg, kidney proximal tubules exhibited progressive degenerative changes from 4 to 24 hr. A functional deficit, decreased uptake of [para-3H]aminohippurate into renal slices, was not observed until 16 hr after i.v. injection of 1 mg/kg. No significant histopathologic changes were observed in kidneys 4 hr after treatment with 0.25 or 0.5 mg Hg/kg, iv. No changes in liver protein synthesis were apparent until 16-24 hr, where an increase in the 70- and 90-kDa proteins was observed. A concomitant increase in plasma sorbitol dehydrogenase activity occurred at 16-24 hr; however, there was no histopathological evidence of liver injury. The 72-kDa inducible member of the 70-kDa stress protein family and the 88-kDa member of the 90-kDa protein family were detected by immunoblotting techniques using monoclonal antibodies. The data demonstrate that Hg induces alterations in the expression of renal gene products in vivo as evidenced by enhanced stress protein synthesis and inhibition of synthesis of constitutive proteins. These changes in renal protein synthesis preceded overt renal injury, occurring in the early stages of nephropathy. Altered patterns of stress protein synthesis appeared to be target organ specific. The data suggest that altered protein synthesis patterns may serve as biomarkers of renal injury.

The heat shock response: potential to screen teratogens

The embryonic stress hypothesis of teratogenesis suggests that a proportion of all human congenital defects is due to a failure in essential gene transcription along with translational pre-emption by the heat shock response (HSR). We sought to determine the potential usefulness of the murine HSR to screen agents suspected of being human teratogens. The teratogenic potential of a selected group of known teratogenic (hyperthermia, insulin, retinoic acid and valproic acid) or non-teratogenic (cycloheximide, dinitrophenol and tetracycline) agents were administered to pregnant SWV mice at critical periods of neural tube closure. Following exposure to either teratogenic doses or at the highest dose possible that did not induce maternal toxicity for those compounds that were not teratogenic, the induction of heat shock protein (hsp) synthesis and changes in total protein synthesis were determined in lymphocytes isolated from murine spleens. The varied results obtained in these studies cast doubt on the value of the murine HSR to screen teratogens.

In vitro analysis of the murine heat shock response: implications for reproductive toxicology

Lymphocytes isolated from two inbred mouse strains that differed in their genetically determined sensitivity to heat-induced exencephaly were used to compare the in vitro kinetics of heat shock protein synthesis in the two strains following hyperthermic exposure. Differences in protein synthesis were determined by densitometric analysis of autoradiograms of SDS-PAGE gels. The findings were consistent with those observed in vivo in that there was an immediate and prolonged synthesis of heat-shock proteins by lymphocytes from the heat sensitive SWV/SD strain, compared to the response observed in lymphocytes from the heat-resistant DBA/2J strain. These results indicate that an in vitro lymphocyte assay of the heat-shock response may be a useful tool for screening suspected teratogenic agents.

Genetic differences in susceptibility to anticonvulsant drug-induced developmental defects

The teratogenicity associated with gestational exposure to anticonvulsant medications remains a significant clinical therapeutic concern. Although most effective antiepileptic agents have been implicated as potential teratogens, the adverse developmental effects associated with phenytoin, trimethadione and valproic acid remain unequivocal. Although approximately 12,000 infants are exposed in utero annually in the United States to these compounds, only 10% will present with the clinical features associated with anticonvulsant drug embryopathies, clearly suggesting the presence of a genetic basis of susceptibility that certain infants inherit. In this article, the genetic basis for such variability in expression, with emphasis on the biochemical and morphological differences that might predominate, are examined in light of recent experimental evidence.

Target tissue specificity of retinoic acid-induced stress proteins and malformations in mice

Retinoic acid (RA) is teratogenic in rodents and also induces the synthesis of stress proteins in fetal mouse limb buds. To determine if the RA induction of stress proteins is target tissue specific, pregnant CD-1 mice were gavaged with 100 mg/kg RA on day 11 of gestation, and nuclei isolated from tissues susceptible to RA-induced malformations (target tissues) as well as nuclei isolated from nontarget tissues were examined for stress protein synthesis and malformations. Forelimb and hindlimb (target tissues), as well as heart and tail (nontarget tissues), were removed from embryos 2.5 hours after RA treatment (1.5 hr after [3H]leucine labeling). Cell nuclei were isolated, stained with a DNA specific fluorochrome, propidium iodide, and sorted from the G0 + G1 and G2 + M phases of the cell cycle. Forelimb and hindlimb target tissues showed the synthesis in these embryonic nuclear proteins of an 84,000 relative molecular mass (Mr) protein and a 90,000 Mr protein following RA treatment. Two 20,000-25,000 Mr stress proteins were also labeled both in forelimb and hindlimb. Forelimb and hindlimb from untreated dams showed no stress protein labeling. Neither heart nor tail, nontarget tissues, showed any stress protein labeling following RA treatment. Classical teratological evaluation of embryos treated on GD 11 and sacrificed on GD 17 showed that 100% of the fetuses had forelimb and/or hindlimb malformations, while no malformations were observed in either the heart or tail. Based on the correlation of teratological anomalies with the identification of stress proteins in target tissue only, we postulate that stress proteins may be involved in the teratogenic process. Further work is necessary to establish whether a causal relationship exists.

Developmental regulation of heat shock protein synthesis and HSP 70 RNA accumulation during postimplantation rat embryogenesis

Exposure of postimplantation rat embryos on days 9, 10, 11, and 12 of gestation to an in vitro heat shock of 43 degrees C for 30 min results in the induction of heat shock proteins (HSPs) in day 9 and 10 embryos, a severely attenuated response in day 11 embryos, and no detectable response in day 12 embryos. The heat shock response in day 9 embryos (presomite stage) is characterized by the synthesis of HSPs with molecular weights of 28-78 kDa. In heat shocked day 10 embryos, two additional HSPs are induced (34 and 82 kDa). In addition, two HSPs present on day 9 are absent on day 10. In day 11 heat shocked embryos, only three HSPs (31, 39, and 69 kDa) are induced, while in day 12 embryos no detectable HSPs are induced. Northern blot analysis of HSP 70 RNA levels indicates that the accumulation of this RNA, but not actin RNA, varies depending on developmental stage at the time of exposure to heat as well as the duration of the heat shock. Day 9 embryos exhibit the most pronounced accumulation of HSP 70 RNA while embryos on days 10-12 exhibit an increasingly attenuated accumulation of HSP 70 RNA, particularly after the more acute exposures (43 degrees C for 30 or 60 min). Thus, the ability to synthesize HSP 70 and to accumulate HSP 70 RNA changes dramatically as rat embryos develop from day 9 to day 12 (presomite to 31-35 somite stages).

Marked increase in the teratogenicity of the combined administration of the industrial solvent 2-methoxyethanol and radiofrequency radiation in rats

Limited published animal research reports synergistic teratogenic effects following combined hyperthermia (induced by elevated ambient temperature) and administration of chemical teratogens. Radiofrequency (RF) radiation is widely used in occupational environments. Since RF radiation also elevates the body temperature of, and is teratogenic to, exposed animals, concurrent RF radiation and chemical agent administration may enhance teratogenicity. The present exploratory study, consisting of preliminary dose-finding studies and the primary study, was designed to investigate whether concurrent exposure of rats to RF radiation and the industrial solvent 2-methoxyethanol (2ME) can enhance the developmental toxicity of either agent acting alone. Preliminary dose-finding studies using small numbers of rats investigated the ability of various RF radiation conditions and doses of 2ME to produce external malformations (primarily of the paws) when administered on gestation day 13. Based on these preliminary studies, RF radiation exposure [sufficient to elevate rectal temperature to 42.0 degrees C (4 degrees C above normal for rats) for 30 min] and 2ME administration (150 mg/kg) were selected for the primary study. In the primary study, groups of 18 to 27 pregnant rats were administered RF radiation exposure and distilled water gavage, 2ME gavage and sham RF exposure, RF radiation exposure and 2ME gavage concurrently, or sham RF exposure and distilled water gavage. Pregnant rats were sacrificed on gestation day 20, and the offspring were examined for external malformations. Combined exposures enhanced the adverse effects produced by either experimental agent alone (no malformations were detected in the double sham group). Mean fetal malformations/litter increased from 14% after 2ME and sham RF (15/26 litters affected, with an average of 2 fetuses/litter malformed) and 30% after RF radiation and water gavage (10/18 litters affected, with an average of 4 fetuses/litter malformed), to 76% after the combined treatment (18/18 litters affected, with an average of 12 fetuses/litter malformed). In addition to a significant increase in the frequency of malformations, the severity of malformations also was enhanced by the combination treatment (on a relative severity ranking scale, the 2ME severity score was less than 1, the RF score was 3, and the combination score was 6). This study provided evidence of synergism between RF radiation and 2ME administration, but additional research will be required to characterize the extent of synergism between these two agents. Potential interactive effects between chemical and physical agents need to be investigated to determine the extent to which such interactions should impact occupational exposure standards.

[Molecular cell biology of the heat stress response. Part I]

In a physiological range of hyperthermia all living systems respond with a complex reprogramming of cellular activities to provide a basis for survival during the stress period and for a rapid restoration of normal activities in the recovery period. A prominent characteristic of the response is the induced synthesis of heat-stress proteins which is likewise evoked by numerous chemical stressors. The common signal transduction chain leading to the activation of heat-stress genes evidently involves the transient accumulation of abnormal proteins. The dominant HSPs belong to five conserved stress protein families, whose members are essential components of all living cells with general functions by far exceeding the stress response.

Effects of heat and chemical stress on development

Similarities in the means by which developmental defects are induced in vertebrates and Drosophila suggest that some kinds of defects may be induced by similar mechanisms. The similarities include the fact that heat and a group of chemicals that induce synthesis of heat-shock proteins induce defects in mammals, chickens, and flies. Different kinds of defects are even produced in one type of animal, depending on the precise timing of the environmental insult. The effectiveness of the environmental treatment in inducing defects depends on the genetic background of the animal as well as on past exposure to chemicals and heat. Developmental defects induced by heat in mice, rats, and flies can all be prevented by thermotolerance-inducing treatments. The basis for these effects has been studied at the molecular level in Drosophila, and the evidence indicates that these teratogens and the thermotolerance-inducing treatments affect the level or timing of expression of specific genes during critical periods in the developmental program.

Embryonic and maternal heat shock responses to a teratogenic hyperthermic insult

Exposing embryos to elevated temperatures both in vivo and in vitro has been shown to result in the production of offspring with severe congenital abnormalities. While a direct effect of heat cannot be excluded, recent interest has been focused on the possible role that the induction of the heat shock response may have in the etiology of the observed congenital defects. In the present study, mouse embryos from inbred strains known to differ in terms of their sensitivity to heat-induced exencephaly were treated in vivo and their heat shock response determined using SDS-PAGE electrophoretic techniques. Further, the embryonic responses were compared with a maternal cell type. We observed excellent agreement between the two test systems following exposure to a teratogenic hyperthermic insult. Both the embryonic and maternal cells underwent a reduction in total protein synthesis and an enhanced synthesis of four heat shock proteins migrating with the molecular weights of 68, 70, 97, and 110 kDa. The results failed to indicate any strong correlation between the heat shock response and enhanced genetic sensitivity to hyperthermia-induced neural tube defects.

Does heat damage fetuses?

Temperature affects phenotypic variation during critical developmental stages in all forms of life that have been studied thus far. In animal studies of heat teratogenicity, adverse effects have ranged from disruption of the normal cell cycle leading to decreased numbers of cells, to the induction of developmental abnormalities by means of embryonic cell death. The heat shock response is a universal cellular stress reaction in which the transcriptional and translational mechanisms of the cell are pre-empted by preferential induction of heat shock protein synthesis. Occurrence of such a phenomenon during prenatal life could lead to the absence of essential gene products at critical stages of development. The crucial question of whether temperature induced cellular and genetic effects ever occur during human fetal development has been considered only in relation to maternal hyperthermia, which is generally viewed as not being of significance in human teratology. We propose that teratogenicity may result from fetal hyperthermia unrelated to maternal hyperthermia, caused either by impaired fetomaternal heat dissipation due to reduced placental blood flow (extrinsic fetal hyperthermia) or by increased fetal heat production during hypermetabolic states (intrinsic fetal hyperthermia). The need for further studies in this regard is emphasized.

Elevated expression of proto-oncogenes accompany enhanced induction of heat-shock genes after exposure of rat embryos in utero to ionizing irradiation

We have recently found that the effects of exposing rat embryos in utero to teratogens capable of producing cardiac anomalies were expressed later as enhanced induction of heat-shock proteins (hsp70 family) when embryonic hearts were cultured in vitro. However, it remained to be determined whether heat-shock proteins are induced in vivo after exposure to teratogens. The heat-shock response in some mammalian systems is known to be accompanied by elevated expression of proto-oncogenes. Using gene-specific DNA probes, we examined the levels of the expression (transcription) of heat-shock protein genes and two nuclear proto-oncogenes, c-fos and c-myc, in the embryos removed from irradiated pregnant mother rats 4 or 5 days after the irradiation. We found that the levels of expression in vivo of the hsp70 and c-myc genes in the irradiated embryos increased by approximately twofold as compared with those in the control. The expression in vivo of the c-fos gene was not detected in either the irradiated or non-irradiated embryos. After 0.5-hr incubation in vitro of the embryos, however, the expression of the c-fos gene in the irradiated embryos was highly enhanced whereas the control showed no changes. Although the exact functions of these gene products still remain obscure, the enhanced expression of hsp70 gene(s) and the nuclear proto-oncogenes observed in the present study may reflect repair of intracellular damages and/or regeneration of tissue by compensatory cell proliferation, processes that may disturb the normal program of organogenesis.

Synthesis and secretion of immunoglobulin G by lymphocytes from cultured mouse spleen cells is not affected by heat shock

Short-term hyperthermic episodes (in vivo and in vitro) alter gene expression in mammalian lymphocytes, resulting in the enhanced synthesis of a select group of polypeptides--the heat-shock proteins--and the depressed synthesis of many normally synthesized polypeptides. Such alterations could have profound implications to an individual if the appropriate functioning of lymphocytes within the immune response was compromised by a depression in immunoglobulin synthesis during naturally occurring periods of hyperthermia, such as fever. In the present study we asked if heat-shock affects the facultative synthesis and secretion of immunoglobulin G by cultured mouse lymphocytes. We found that the quantity of immunoglobulin G synthesized and secreted by these cells is not affected by heat-shock treatments sufficient to induce the synthesis of heat-shock proteins.

The effects of heat shock on the morphology and protein synthesis of the epidermis of Xenopus laevis larvae

By scanning electron microscopy, we have observed that a 20-min heat shock at 37 degrees C, although not lethal, causes extensive damage to the epidermis of 30-h and 2-d (post-fertilization) Xenopus laevis larvae. The primary effects of heat shock are the apical swelling of the epidermal cells, giving the epidermis a "cobblestone" appearance, and the selective shedding of the ciliated cells. The shed cells may be cell fragments, however, because some of them are anucleate. Shed cells also exhibit the enriched synthesis of a group of heat shock proteins of 62,000 D molecular weight, suggesting that these proteins are specific to the shed cells. Prolonged heat shock of these larvae (i.e., 30 min at 37 degrees C) results in the complete disintegration of the epidermis, followed by larval death. At later stages of development (3-d and 4-d post-fertilization), the epidermis becomes more resistant to heat-induced damage inflicted by a 20-min heat shock. This increase in resistance coincides with the development of large secretory cells and the loss of ciliated cells in the epidermis and thus parallels a change in the state of histological differentiation.

Effects of exposing rat embryos in utero to physical or chemical teratogens are expressed later as enhanced induction of heat-shock proteins when embryonic hearts are cultured in vitro

In order to get more insight into the effects of teratogens on developing embryos, we investigated the protein synthesis patterns of the target organs isolated from teratogen-treated embryos. Rat embryos were either irradiated in utero with either 252Cf fission neutrons or 60Co gamma rays on day 8 of gestation or treated in utero with a bis(dichloroacetyl)diamine (a chemical teratogen) on days 9 and 10. Hearts were removed from the embryos on day 12 and were incubated in vitro at 37 degrees C in the presence of [35S]methionine for up to 8 hr. The newly synthesized labeled proteins were then analyzed qualitatively by two-dimensional polyacrylamide gel electrophoresis. Enhanced and prolonged induction of a family of heat-shock (stress) proteins with a molecular weight of about 70,000 (SP70s) was observed as compared with those of controls. Among the teratogen-treated hearts, those with gross malformations already detectable at this early stage showed especially higher inductions of SP70s than did the others. The abnormal expression of SP70s observed in the present study appears to be a reflection of persisting cellular (tissue) damage inflicted by the teratogens, and the extent of the induction may be indicative of the degree and/or type of the damage. Such persisting defects in surviving cells, manifested by abnormal induction of SP70s in the present study, might be related to malformation of embryonic hearts.

Heat shock and thermotolerance during early rat embryo development

Effects of heat shock on the development of early pre-somite embryos have been studied using cultured rat embryos. The results illustrate the sensitivity of the developing head and brain to elevated temperatures prior to neural tube closure and the capacity of embryos to acquire thermotolerance. Embryos exposed briefly to an elevated temperature (43 degrees C for 7.5 min) developed severe craniofacial defects including microphthalmia, microcephaly, gross reduction of the forebrain region, and open neural tubes. In contrast, a nonteratogenic heat shock (42 degrees C for 10 min) caused embryos to acquire thermotolerance during a 15-min recovery period at 38.5 degrees C. Acquired thermotolerance was effective in protecting embryos from a subsequent more severe heat treatment which would have been teratogenic in an unprotected embryo. Recovering embryos mounted a heat shock response as evidenced by the induction of a 71 kilodalton heat shock protein. Activation of the heat shock response was not a teratogenic event in the developing embryo.

Heat-shock proteins and development

At the simplest level there is little doubt that the heat shock response is homeostatic, to protect the cell against the ravages of the environmental insult and ensure that the cell can continue its normal life after the crisis has passed.

The potential relationship of maternal toxicity, general stress, and fetal outcome

Standard teratology bioassays generally call for a top dose level which is sufficient to induce some form of overt maternal toxicity such as death or weight loss. The presence of such maternal toxicity is often a confound in the interpretation of experimental results, especially at those dose levels producing the toxicity. While the physiological bases for the toxicity vary widely in a compound-related fashion, one underlying factor that remains constant for most induced toxicity is the presence of generalized stress in the affected animals. Previous studies have indicated that pregnant animals treated acutely with toxic levels of a variety of pharmacologically unrelated chemicals produced litters without a recognizable syndrome of defects, except for an increased incidence of supernumerary ribs (SNR). The present study reports on the effects of immobilization stress on the production of SNR in the Sprague-Dawley rat and the CD-1 mouse. Pregnant animals were immobilized in the supine position for 12-hour periods during the day of greatest sensitivity to SNR production (days 9 and 10 in the mouse and rat, respectively). Animals were killed immediately before term and the fetuses were examined. An increase in SNR was noted in immobilized mice but not rats. These results suggest that such fetal effects may be the result of general agent-induced maternal stress.

Hyperthermia-induced heat shock response and thermotolerance in postimplantation rat embryos

Postimplantation stage rat embryos (6-10 somites) undergo abnormal development after exposure to a temperature of 43 degrees C for 30 min. A heat shock of 43 degrees C for 30 min also induces the synthesis of a set of eight heat shock proteins (hsps) with molecular masses ranging from 28,000 to 82,000 Da. The synthesis of these hsps is rapidly induced after the heat shock is applied and rapidly decays after embryos are returned to 37 degrees C. A heat shock of 42 degrees C for 30 min has no effect on rat embryo growth and development, but does induce the synthesis of three hsps. The most prominent of these three is believed to be the typical mammalian 70 kDa hsp. Furthermore, a 42 degrees C, 30-min heat shock followed by a 43 degrees C 30-min heat shock leads to partial protection from the embryotoxic effects of a single exposure at 43 degrees C, i.e., thermotolerance.

Hyperthermia as a teratogen: a review of experimental studies and their clinical significance

Although hyperthermia is teratogenic in birds, all the common laboratory animals, farm animals, and primates and satisfies defined criteria as a teratogen, its study as a human teratogen has been neglected. Homeothermic animals, including humans, can experience body temperature elevations induced by febrile infections, heavy exercise and hot environments which exceed the thresholds (1.5-2.5 degrees C elevation) which are known to cause a syndrome of embryonic resorptions, abortions, and malformations in experimental animals. Hyperthermia is particularly damaging to the central nervous system, and if a threshold exposure occurs at the appropriate stages of embryonic development, exencephaly, anencephaly, encephalocoele, micrencephaly, microphthalmia, neurogenic talipes, and arthrogryposis can be produced in a high proportion of exposed embryos, the incidence and type of defect depending on the species and strain within species, the stage of development, and the severity of hyperthermic exposure. Other defects which can be induced experimentally include exomphalos, hypoplasia of toes and teeth, renal agenesis, vertebral anomalies, maxillary hypoplasia, facial clefting, cataract, coloboma, and heart and vascular defects. Proliferating cells are particularly sensitive to temperature elevations, resulting in arrest of mitotic activity and immediate death of cells in mitosis with threshold elevations (1.5-2.5 degrees C) and delayed death of cells probably in S phase with higher elevations (3.5 degrees C). In general, lower temperature elevations (2.5 degrees C) require longer durations of elevation to cause defects than a simple spike at a higher elevation (4.5 degrees C). The death of cells is largely confined to the brain and in the day 21 guinea pig embryo to the alar regions of the brain. Cell death probably accounts for most of the defects in the central nervous system, but microvascular disturbances leading to leakage, oedema and haemorrhage, placental necrosis, and infarction are other known effects of hyperthermia; and these are probably involved in the pathogenesis of many defects of the heart, limbs, kidneys, and body wall. Recent experiments have demonstrated protection of rat embryos in culture against a known teratogenic exposure by a brief nonteratogenic exposure given at least 15 min earlier. This protection is associated with the synthesis of heat-shock proteins, and temporary arrest of the cell proliferative cycle. Hyperthermia appears to be capable of causing congenital defects in all species and may act alone or synergistically with other agents.(ABSTRACT TRUNCATED AT 400 WORDS)

The induction of supernumerary ribs in rodents: role of the maternal stress

Increased incidences of supernumerary ribs (SNR) are a relatively common finding in standard teratology bioassays, and previous studies have indicated a possible correlation between their occurrence and general maternal stress. The present study describes the effects of immobilization stress on the induction of supernumerary ribs. To isolate the sensitive period of SNR induction, Sprague-Dawley rats and CD-1 mice were treated with 300 and 1,500 mg/kg, respectively, of sodium salicylate on single days 7-11 of gestation. In the rat, day 10 was found to be the sensitive period of lumbar rib induction (32% SNR vs 10% on other days) while day 9 was critical in the mouse (71% SNR vs less than 29% on other days). In a second set of experiments, maternal stress was accomplished by restraining two groups of gravid females in the supine position for 12 hours on the predetermined sensitive day. One group was immobilized from 9 am to 9 pm, while the second group was restrained from 9 pm to 9 am. Concurrent controls were food and water deprived for similar periods. Additional untreated controls were also included. An increase in supernumerary ribs was noted in stressed mice but not in rats. The 9 am to 9 pm mouse group exhibited the highest increase in supernumerary ribs (41%) as well as significant incidences of fused ribs and exencephaly. A significant linear relationship between maternal weight loss during treatment and increases in supernumerary ribs was also noted. Results suggest that for some compounds, supernumerary ribs may be the indirect result of agent-induced, generalized maternal stress in the CD-1 mouse.

The heat-shock response in vivo: experimental induction during mammalian organogenesis

According to the embryonic stress hypothesis of teratogenesis, anatomical malformation can be the consequence of the induction of a heat-shock response (HSR) in the embryo at some critical stage during the determination or differentiation of organs. This hypothesis states that a teratogen is any agent that is capable of inducing a HSR and that can reach the developing embryo. As a first step in determining whether the hypothesis is tenable, it was necessary to determine whether the embryo in fact is capable of making the HSR during the period of organogenesis. Pregnant mice were treated with two classical inducers of the HSR, one a physical and the other a chemical agent--namely, hyperthermia and sodium arsenite. The embryos, while still in the living mouse, responded with heat-shock protein induction, as did control bone marrow.