The evolving microenvironment of the human hepatocyte: Healthy vs. cirrhotic liver vs. isolated cells

The study of the liver microenvironment and hepatocyte's response to this environment in the setting of healthy liver, cirrhotic liver or cultured primary human hepatocytes (PHHs) addresses key questions for the development of novel liver therapies and predicts relevance of ex vivo PHHs models in liver biology. This study compared quantitative gene and protein expression of the inflammatory profile, oxidative stress response, angiogenesis and homing mechanisms in the biopsies of healthy and cirrhotic human livers and isolated PHHs. These profiles were correlated with the metabolic health of liver and PHHs defined by albumin production. The analysis demonstrated that cirrhotic liver and PHHs exhibited a distinct upregulation of the pro-inflammatory, oxidative stress and homing mechanism markers when compared to normal liver. The upregulation of the oxidative stress markers in PHHs inversely correlated with the albumin production. PHHs had diverse secretion of matrix metalloproteinases and their inhibitors, reflective of the cellular response to non-physiological culture conditions. The current study suggests that ex vivo PHHs manifest adaptive behavior by upregulating stress mechanisms (similar to the cirrhotic liver), downregulating normal metabolic function and upregulating matrix turnover. The ex vivo profile of PHHs may limit their therapeutic functionality and metabolic capacity to serve as in vitro metabolism and toxicology models.

Hemodynamic flow improves rat hepatocyte morphology, function, and metabolic activity in vitro

In vitro primary hepatocyte systems typically elicit drug induction and toxicity responses at concentrations much higher than corresponding in vivo or clinical plasma C(max) levels, contributing to poor in vitro-in vivo correlations. This may be partly due to the absence of physiological parameters that maintain metabolic phenotype in vivo. We hypothesized that restoring hemodynamics and media transport would improve hepatocyte architecture and metabolic function in vitro compared with nonflow cultures. Rat hepatocytes were cultured for 2 wk either in nonflow collagen gel sandwiches with 48-h media changes or under controlled hemodynamics mimicking sinusoidal circulation within a perfused Transwell device. Phenotypic, functional, and metabolic parameters were assessed at multiple times. Hepatocytes in the devices exhibited polarized morphology, retention of differentiation markers [E-cadherin and hepatocyte nuclear factor-4α (HNF-4α)], the canalicular transporter [multidrug-resistant protein-2 (Mrp-2)], and significantly higher levels of liver function compared with nonflow cultures over 2 wk (albumin ~4-fold and urea ~5-fold). Gene expression of cytochrome P450 (CYP) enzymes was significantly higher (fold increase over nonflow: CYP1A1: 53.5 ± 10.3; CYP1A2: 64.0 ± 15.1; CYP2B1: 15.2 ± 2.9; CYP2B2: 2.7 ± 0.8; CYP3A2: 4.0 ± 1.4) and translated to significantly higher basal enzyme activity (device vs. nonflow: CYP1A: 6.26 ± 2.41 vs. 0.42 ± 0.015; CYP1B: 3.47 ± 1.66 vs. 0.4 ± 0.09; CYP3A: 11.65 ± 4.70 vs. 2.43 ± 0.56) while retaining inducibility by 3-methylcholanthrene and dexamethasone (fold increase over DMSO: CYP1A = 27.33 and CYP3A = 4.94). These responses were observed at concentrations closer to plasma levels documented in vivo in rats. The retention of in vivo-like hepatocyte phenotype and metabolic function coupled with drug response at more physiological concentrations emphasizes the importance of restoring in vivo physiological transport parameters in vitro.

Regulation of cell morphology and cytochrome P450 expression in human hepatocytes by extracellular matrix and cell-cell interactions

The influence of extracellular matrix conditions and plating density on cell cytoarchitecture and the constitutive and chemically induced expression of cytochrome P450 3A4 (CYP3A4) was examined in primary cultures of human hepatocytes. Constitutive and drug-induced microsomal CYP3A4 expression occurred equally well in human hepatocyte cultures maintained on either a complex or simple substratum (Matrigel vs collagen, type I), or in a sandwich configuration (i.e., between two layers of extracellular matrix), despite the markedly different morphological properties exhibited by each condition. However, a density-dependent decrease in both the constitutive and induced levels of CYP3A4 was observed in hepatocytes maintained on a simple collagen substratum as plating density was reduced from 100% to 25%. Marked alterations in cell shape and cytoarchitecture were noted concomitant with decreases in the expression and localization of intercellular gap junctions and E-cadherin-mediated cell adhesions. In addition, the intracellular distribution of microtubules and microfilaments was altered substantially and the expression of immunoreactive actin and beta-tubulin increased as cell density was decreased. These effects were reversed to some extent by overlaying monolayers with extracellular matrix or by co-culturing with another cell type. Efforts to maintain normal cell shape and cytoskeletal distribution in hepatocytes at low cell density with a Matrigel substratum failed to restore normal basal levels of CYP3A4 expression or responsiveness to rifampicin (RIF). Likewise, E-cadherin and Cx-32 expression was again reduced, even though the distribution and expression of cytoskeletal elements returned to normal levels. These results suggest that cell-cell contacts, but not the extracellular matrix configuration or composition, play a critical role in determining normal responsiveness to chemical modulators in human hepatocytes.

Human hepatocyte culture systems for the in vitro evaluation of cytochrome P450 expression and regulation

Primary cultures of human hepatocytes have been used extensively by both academic and industrial laboratories for evaluating the hepatic disposition of drugs and other xenobiotics. Their primary utility has been for assessing the induction potential of new chemical entities (NCEs) and they continue to serve as the gold standard. Primary considerations for conducting in vitro drug testing utilizing cultures of human hepatocytes, such as the effects of culture and study conditions, are discussed. The maintenance of normal cellular physiology and intercellular contacts in vitro is of particular importance for optimal phenotypic gene expression and response to drugs and other xenobiotics. Significant advances in our understanding of cytochrome P450 (CYP450) enzyme regulation have been made with the recent identification of the nuclear receptors mediating the induction of CYP2B and CYP3A enzymes. In particular, the activation of pregnane X receptor (PXR) by prototypical inducers of CYP3A has been found to correlate well with the species-specific modulation of CYP3A by various drugs and other xenobiotics. Concomitant with the discovery of PXR has been the identification of compounds that may act synergistically or antagonistically on multiple receptors (e.g., co-repressors and/or co-activators of the receptor) introducing novel mechanisms of drug-drug interactions. Differential expression of the individual isoforms of the major CYP450 enzymes over time in culture suggest that this model system is not reflective of in vivo profiles and, therefore, may be limited in its application for drug metabolism studies. Overall, primary cultures of human hepatocytes can serve as a sensitive and selective model for predicting the regulation of CYP450 modulation by drugs and other xenobiotics. Considerations and recommendations for standardizing testing conditions and choosing relevant endpoint(s) are presented.

Pregnane X receptor: molecular basis for species differences in CYP3A induction by xenobiotics

Determining the molecular basis for the observed species differences in the xenobiotic-mediated induction of cytochrome 3A (CYP3A) gene expression has become one of the biggest dilemmas of the modern era in toxicology. Recently, a novel orphan nuclear receptor, termed pregnane X receptor (PXR), has been implicated to play a key role in the regulation of CYP3A genes by xenobiotics. PXR is capable of binding to and activating transcription from specific response elements found in the CYP3A gene promoter from multiple species. Notably, compounds that are known to induce CYP3A selectively in human, mouse, rat, or rabbit also activate the corresponding PXR. Pregnenolone 16alpha-carbonitrile, a known CYP3A inducer in rodents, is a very efficacious activator of mouse and rat PXR, whereas rifampicin, a known inducer of CYP3A in humans and rabbits, is a very efficacious activator of human and rabbit PXR. Likewise, selective activators of PXR also induce CYP3A gene expression in the corresponding species. Orthologous receptors from human, mouse, rat, and rabbit have been cloned and characterized and share approx. 95% identity in their DNA binding domains. By contrast, they share only 75-80% identity in their amino acid sequences in the ligand-binding domain. Together, these data suggest that PXR is a critical regulator of CYP3A gene expression and activation of PXR is predictive of CYP3A induction. Furthermore, sequence differences in the ligand-binding domain, and not the DNA binding domain, appear to serve as the molecular basis for the species differences in CYP3A induction observed in vivo.

Pregnane X receptor: molecular basis for species differences in CYP3A induction by xenobiotics

Determining the molecular basis for the observed species differences in the xenobiotic-mediated induction of cytochrome 3A (CYP3A) gene expression has become one of the biggest dilemmas of the modern era in toxicology. Recently, a novel orphan nuclear receptor, termed pregnane X receptor (PXR), has been implicated to play a key role in the regulation of CYP3A genes by xenobiotics. PXR is capable of binding to and activating transcription from specific response elements found in the CYP3A gene promoter from multiple species. Notably, compounds that are known to induce CYP3A selectively in human, mouse, rat, or rabbit also activate the corresponding PXR. Pregnenolone 16alpha-carbonitrile, a known CYP3A inducer in rodents, is a very efficacious activator of mouse and rat PXR, whereas rifampicin, a known inducer of CYP3A in humans and rabbits, is a very efficacious activator of human and rabbit PXR. Likewise, selective activators of PXR also induce CYP3A gene expression in the corresponding species. Orthologous receptors from human, mouse, rat, and rabbit have been cloned and characterized and share approx. 95% identity in their DNA binding domains. By contrast, they share only 75-80% identity in their amino acid sequences in the ligand-binding domain. Together, these data suggest that PXR is a critical regulator of CYP3A gene expression and activation of PXR is predictive of CYP3A induction. Furthermore, sequence differences in the ligand-binding domain, and not the DNA binding domain, appear to serve as the molecular basis for the species differences in CYP3A induction observed in vivo.

Application of cDNA microarray to the study of arsenic-induced liver diseases in the population of Guizhou, China

Arsenic is an environmental toxicant and a human carcinogen. Epidemiology studies link human arsenic exposure to various diseases and cancers, including liver diseases and hepatocellular carcinoma. However, the molecular mechanisms for arsenic toxicity and carcinogenicity are poorly understood. To better understand these mechanisms, we used the human cancer cDNA expression array to profile aberrant gene expression in arsenic-exposed populations in Guizhou, China. The selected patients had a history of exposure to environmental arsenic for at least 6-10 years, and had arsenic-induced skin lesions and hepatomegaly. Samples were obtained by liver needle biopsy. Histology showed degenerative liver lesions, such as chronic inflammation, vacuolation, and focal necrosis. The University of North Carolina Hospitals provided normal human liver tissues from surgical resection or rejected transplants. Microarray was performed with total RNA from liver samples, and signal intensities were analyzed with AtlasImage software and normalized with 9 housekeeping genes. Means and SEM were calculated for statistical analysis. Approximately 60 genes (10%) were differentially expressed in arsenic-exposed human livers compared to controls. The differentially expressed genes included those involved in cell-cycle regulation, apoptosis, DNA damage response, and intermediate filaments. The observed gene alterations appear to be reflective of hepatic degenerative lesions seen in the arsenic-exposed patients. This array analysis revealed important patterns of aberrant gene expression occurring with arsenic exposure in human livers. Aberrant expressions of several genes were consistent with the results of array analysis of chronic arsenic-exposed mouse livers and chronic arsenic-transformed rat liver cells. Clearly, a variety of gene expression changes may play an integral role in arsenic hepatotoxicity and possibly carcinogenesis.

In vivo and in vitro induction of human cytochrome P4503A4 by dexamethasone

PURPOSE

The aims of these experiments were to determine the effect of a therapeutic regimen of dexamethasone on cytochrome P4503A4 (CYP3A4) activity in healthy volunteers; and the concentration-effect relationship between dexamethasone and CYP3A4 activity in primary human hepatocyte cultures.

METHODS

The effect of dexamethasone (8 mg administered by mouth two times a day for 5 days) on CYP3A4 activity in 12 healthy volunteers was assessed with the erythromycin breath test and urinary ratio of dextromethorphan to 3-methoxymorphinan. Concentration-effect of dexamethasone on CYP3A4-dependent testosterone 6-beta-hydroxylation was determined in human hepatocytes treated with 2 to 250 micromol/L dexamethasone.

RESULTS

The percent of erythromycin metabolized per hour increased from 2.20% +/- 0.60% (mean +/- SD) at baseline to 2.67% +/- 0.55% on day 5 of dexamethasone (mean increase in hepatic CYP3A4 activity 25.7% +/- 24.6%; P = .004). The mean urinary ratio of dextromethorphan to 3-methoxymorphinan was 28 (4.8 to 109) and 7 (1 to 23) at baseline and on day 5 of dexamethasone (mean decrease = 49%; P = .06). Substantial intersubject variability was observed in the extent of CYP3A4 induction. The extent of CYP3A4 induction was inversely correlated with baseline erythromycin breath test (r2 = 0.58). In hepatocytes, dexamethasone 2 to 250 micromol/L resulted in an average 1.7-fold to 6.9-fold increase in CYP3A4 activity, respectively. The extent of CYP3A4 induction with dexamethasone in hepatocyte preparations was inversely correlated with baseline activity (r2 = 0.59).

CONCLUSIONS

These data demonstrate that dexamethasone at doses used clinically increased CYP3A4 activity with extensive intersubject variability and that the extent of CYP3A4 induction was, in part, predicted by the baseline activity of CYP3A4 in both healthy volunteers and human hepatocyte cultures.

Comparative toxicity of trivalent and pentavalent inorganic and methylated arsenicals in rat and human cells

Biomethylation is considered a major detoxification pathway for inorganic arsenicals (iAs). According to the postulated metabolic scheme, the methylation of iAs yields methylated metabolites in which arsenic is present in both pentavalent and trivalent forms. Pentavalent mono- and dimethylated arsenicals are less acutely toxic than iAs. However, little is known about the toxicity of trivalent methylated species. In the work reported here the toxicities of iAs and trivalent and pentavalent methylated arsenicals were examined in cultured human cells derived from tissues that are considered a major site for iAs methylation (liver) or targets for carcinogenic effects associated with exposure to iAs (skin, urinary bladder, and lung). To characterize the role of methylation in the protection against toxicity of arsenicals, the capacities of cells to produce methylated metabolites were also examined. In addition to human cells, primary rat hepatocytes were used as methylating controls. Among the arsenicals examined, trivalent monomethylated species were the most cytotoxic in all cell types. Trivalent dimethylated arsenicals were at least as cytotoxic as trivalent iAs (arsenite) for most cell types. Pentavalent arsenicals were significantly less cytotoxic than their trivalent analogs. Among the cell types examined, primary rat hepatocytes exhibited the greatest methylation capacity for iAs followed by primary human hepatocytes, epidermal keratinocytes, and bronchial epithelial cells. Cells derived from human bladder did not methylate iAs. There was no apparent correlation between susceptibility of cells to arsenic toxicity and their capacity to methylate iAs. These results suggest that (1) trivalent methylated arsenicals, intermediary products of arsenic methylation, may significantly contribute to the adverse effects associated with exposure to iAs, and (2) high methylation capacity does not protect cells from the acute toxicity of trivalent arsenicals.

Comparative toxicity of trivalent and pentavalent inorganic and methylated arsenicals in rat and human cells

Biomethylation is considered a major detoxification pathway for inorganic arsenicals (iAs). According to the postulated metabolic scheme, the methylation of iAs yields methylated metabolites in which arsenic is present in both pentavalent and trivalent forms. Pentavalent mono- and dimethylated arsenicals are less acutely toxic than iAs. However, little is known about the toxicity of trivalent methylated species. In the work reported here the toxicities of iAs and trivalent and pentavalent methylated arsenicals were examined in cultured human cells derived from tissues that are considered a major site for iAs methylation (liver) or targets for carcinogenic effects associated with exposure to iAs (skin, urinary bladder, and lung). To characterize the role of methylation in the protection against toxicity of arsenicals, the capacities of cells to produce methylated metabolites were also examined. In addition to human cells, primary rat hepatocytes were used as methylating controls. Among the arsenicals examined, trivalent monomethylated species were the most cytotoxic in all cell types. Trivalent dimethylated arsenicals were at least as cytotoxic as trivalent iAs (arsenite) for most cell types. Pentavalent arsenicals were significantly less cytotoxic than their trivalent analogs. Among the cell types examined, primary rat hepatocytes exhibited the greatest methylation capacity for iAs followed by primary human hepatocytes, epidermal keratinocytes, and bronchial epithelial cells. Cells derived from human bladder did not methylate iAs. There was no apparent correlation between susceptibility of cells to arsenic toxicity and their capacity to methylate iAs. These results suggest that (1) trivalent methylated arsenicals, intermediary products of arsenic methylation, may significantly contribute to the adverse effects associated with exposure to iAs, and (2) high methylation capacity does not protect cells from the acute toxicity of trivalent arsenicals.

Regulation of glutathione S-transferase enzymes in primary cultures of rat hepatocytes maintained under various matrix configurations

Primary rat hepatocytes were cultured under various matrix and media conditions and examined after 1 week for the expression and regulation of cytosolic glutathione S-transferase (GST) enzymes. Striking effects on cell morphology were observed in relation to the different matrix conditions, whereas media effects were less prominent. Hepatocytes cultured in serum-free Dulbecco's modified Eagle's medium (DMEM) or modified Chee's medium (MCM) maintained similar levels of total GST protein regardless of the matrix configuration or corresponding cell integrity. However, HPLC analysis showed a differential expression pattern of individual GST subunits in both a time- and medium-dependent fashion. A variable, but pronounced, matrix and medium effect was observed on the induction of total GST expression by various prototypical inducers. Dexamethasone (10 microM) induced subunits A2, M1 and M2 in a medium- and matrix-dependent fashion, whereas phenobarbital (100 microM) induced significantly only subunit A2. beta-Naphthoflavone (50 microM) suppressed all GST subunit expression except subunit P1, which was induced in a matrix- and medium-dependent fashion. These studies show that total basal level expression of GSTs in vitro is reflective of a concomitant increase in mu and pi class subunits and a decrease in alpha class subunits. Moreover, the matrix and medium conditions influence both the basal and inducible expression of GST subunits in cultured rat hepatocytes.

Regeneration and maintenance of bile canalicular networks in collagen-sandwiched hepatocytes

The morphological and cytoskeletal reorganization of collagen-sandwiched rat hepatocytes during the de novo formation of complete canalicular networks was examined by phase, fluorescence and electron microscopy. During the initial stages of membrane repolarization, there was a marked accumulation of both microfilaments and microtubules at the sites of canalicular generation. Microtubule-disrupting agents (colchicine, nocodazole) inhibited the localization of actin filaments at cell margins and the initiation and branching of canalicular networks. After removal of microtubule-disrupting agents, microfilaments relocalized to the canalicular borders and microtubules nucleated along the margins of the bile canaliculi at sites distinct from the peri-canalicular actin networks. Microfilament-perturbing agents (cytochalasin D, phalloidin) did not affect the de novo initiation of bile canaliculi and only slightly impaired the development of canalicular lumina into networks. In established cultures with complete canalicular networks, subsequent treatment with microtubule-disrupting agents did not acutely affect the integrity of preformed canalicular networks. In contrast, treatment with microfilament-perturbing agents caused a marked dilation of most canaliculi. These results illustrate the differential role of the cytoskeleton in the regeneration and maintenance of bile canalicular networks by collagen-sandwiched hepatocytes. Moreover, this study shows the utility of this system as an in vitro model for examining the regulation of cell and membrane polarity.

Effect of troglitazone on cytochrome P450 enzymes in primary cultures of human and rat hepatocytes

1. Troglitazone was the first thiazolidinedione approved for clinical use in the treatment of non-insulin-dependent diabetes mellitus. During clinical investigations of drug-drug interactions with therapeutics (terfenadine and cyclosporine) known to be metabolized by CYP3A4, pharmacokinetic interactions were noted upon troglitazone multiple-dose treatments. The nature of the interactions suggested induction of CYP3A enzymes. 2. Primary cultures of human hepatocytes were used to investigate the induction potential of troglitazone with respect to CYP3A4, CYP2B6 and CYP1A1/2. In human hepatocytes, troglitazone induced both immunoreactive CYP3A4 protein and testosterone 6beta-hydroxylase activity in a dose-dependent fashion (EC50 = 5-10 microM), accompanied by an increase in CYP3A4 mRNA. The capacity of troglitazone to induce CYP3A4 was between that of rifampin (EC50 = 0.8 microM) and dexamethasone (40-50 microM). Troglitazone increased CYP2B6 immunoreactive protein but did not significantly effect CYP1A1/2 activity, immunoreactive protein or mRNA. 3. Troglitazone produced significant increases in CYP3A message, protein and activity in primary rat hepatocytes, a slight increase in CYP2B1/2 activity and no change in CYP1A1/2 message or activity. 4. These results provide evidence that troglitazone can induce CYP3A and CYP2B enzymes while apparently not altering CYP1A. This provides a rationale for the clinically observed interactions of troglitazone with selected CYP3A4 substrates.

The pregnane X receptor: a promiscuous xenobiotic receptor that has diverged during evolution

Transcription of genes encoding cytochrome P450 3A (CYP3A) monooxygenases is induced by a variety of xenobiotics and natural steroids. There are marked differences in the compounds that induce CYP3A gene expression between species. Recently, the mouse and human pregnane X receptor (PXR) were shown to be activated by compounds that induce CYP3A expression. However, most studies of CYP3A regulation have been performed using rabbit and rat hepatocytes. Here, we report the cloning and characterization of PXR from these two species. PXR is remarkably divergent between species, with the rabbit, rat, and human receptors sharing only approximately 80% amino acid identity in their ligand-binding domains. This sequence divergence is reflected by marked pharmacological differences in PXR activation profiles. For example, the macrolide antibiotic rifampicin, the antidiabetic drug troglitazone, and the hypocholesterolemic drug SR12813 are efficacious activators of the human and rabbit PXR but have little activity on the rat and mouse PXR. Conversely, pregnane 16alpha-carbonitrile is a more potent activator of the rat and mouse PXR than the human and rabbit receptor. The activities of xenobiotics in PXR activation assays correlate well with their ability to induce CYP3A expression in primary hepatocytes. Through the use of a novel scintillation proximity binding assay, we demonstrate that many of the compounds that induce CYP3A expression bind directly to human PXR. These data establish PXR as a promiscuous xenobiotic receptor that has diverged during evolution.

The pregnane X receptor: a promiscuous xenobiotic receptor that has diverged during evolution

Transcription of genes encoding cytochrome P450 3A (CYP3A) monooxygenases is induced by a variety of xenobiotics and natural steroids. There are marked differences in the compounds that induce CYP3A gene expression between species. Recently, the mouse and human pregnane X receptor (PXR) were shown to be activated by compounds that induce CYP3A expression. However, most studies of CYP3A regulation have been performed using rabbit and rat hepatocytes. Here, we report the cloning and characterization of PXR from these two species. PXR is remarkably divergent between species, with the rabbit, rat, and human receptors sharing only approximately 80% amino acid identity in their ligand-binding domains. This sequence divergence is reflected by marked pharmacological differences in PXR activation profiles. For example, the macrolide antibiotic rifampicin, the antidiabetic drug troglitazone, and the hypocholesterolemic drug SR12813 are efficacious activators of the human and rabbit PXR but have little activity on the rat and mouse PXR. Conversely, pregnane 16alpha-carbonitrile is a more potent activator of the rat and mouse PXR than the human and rabbit receptor. The activities of xenobiotics in PXR activation assays correlate well with their ability to induce CYP3A expression in primary hepatocytes. Through the use of a novel scintillation proximity binding assay, we demonstrate that many of the compounds that induce CYP3A expression bind directly to human PXR. These data establish PXR as a promiscuous xenobiotic receptor that has diverged during evolution.

Dodecylphosphocholine-mediated enhancement of paracellular permeability and cytotoxicity in Caco-2 cell monolayers

The intestinal epithelium is a significant barrier for oral absorption of hydrophilic drugs because they cannot easily traverse the lipid bilayer of the cell membrane and their passage through the intercellular space (paracellular transport) is restricted by the tight junctions. In this report we show that dodecylphosphocholine (DPC) can improve the paracellular permeability of hydrophilic compounds across Caco-2 cell monolayers by modulating the tight junctions. The results show that the alkyl chain as well as the zwitterionic head group of DPC are required for its activity. DPC appears to act by modulating the permeability of tight junctions as evidenced by the fact that treatment of Caco-2 cell monolayers by this agent results in a decreased transepithelial electrical resistance (TEER), increased permeability of paracellular markers (e. g., mannitol) with no change in the permeability of the transcellular marker testosterone, and redistribution of the tight junction-associated protein ZO-1. The effect of DPC on Caco-2 cells (e.g., decrease in TEER) is reversible, and is not caused by gross cytotoxicity (as indicated by the MTT test) or by nonspecific disruption of the cell membrane (as indicated by only slight nuclear staining due to the nonpermeable DNA-specific dye propidium iodide). We propose in the present study a parameter, potency index, that allows comparison of various enhancers of paracellular transport in relation to their cytotoxicity. The potency index is a ratio between the IC(50) value (concentration at which 50% inhibition of control mitochondrial dehydrogenase activity occurs in the MTT test) and the EC(50) value (concentration at which TEER drops to 50% of its control (untreated) value). By this parameter, DPC is significantly safer than the commonly used absorption enhancer palmitoyl carnitine (PC), which has the potency index of approximately 1 (i.e., no separation between effective and toxic concentration).

Metabolism of arsenic in primary cultures of human and rat hepatocytes

The liver is considered a major site for methylation of inorganic arsenic (iAs). However, there is little data on the capacity of human liver to methylate iAs. This work examined the metabolism of arsenite (iAs(III)), arsenate (iAs(V)), methylarsine oxide (MAs(III)O), methylarsonic acid (MAs(V)), dimethylarsinous acid (DMAs(III)), and dimethylarsinic acid (DMAs(V)) in primary cultures of normal human hepatocytes. Primary rat hepatocytes were used as methylating controls. iAs(III) and MAs(III)O were metabolized more extensively than iAs(V) and MAs(V) by either cell type. Neither human nor rat hepatocytes metabolized DMAs(III) or DMAs(V). Methylation of iAs(III) by human hepatocytes yielded methylarsenic (MAs) and dimethylarsenic (DMAs) species; MAs(III)O was converted to DMAs. The total methylation yield (MAs and DMAs) increased over the range of 0.1 to 4 microM iAs(III). However, DMAs production was inhibited by iAs(III) in a concentration-dependent manner, and the DMAs/MAs ratio decreased. iAs(III) (10 and 20 microM) inhibited both methylation reactions. Inhibition of DMAs synthesis resulted in accumulation of iAs and MAs in human hepatocytes, suggesting that dimethylation is required for iAs clearance from cells. Methylation capacities of human hepatocytes obtained from four donors ranged from 3.1 to 35.7 pmol of iAs(III) per 10(6) cells per hour and were substantially lower than the methylation capacity of rat hepatocytes (387 pmol of iAs(III) per 10(6) cells per hour). The maximal methylation rates for either rat or human hepatocytes were attained between 0.4 and 4 microM iAs(III). In summary, (i) human hepatocytes methylate iAs, (ii) the capacities for iAs methylation vary among individuals and are saturable, and (iii) moderate concentrations of iAs inhibit DMAs synthesis, resulting in an accumulation of iAs and MAs in cells.

Metabolism of arsenic in primary cultures of human and rat hepatocytes

The liver is considered a major site for methylation of inorganic arsenic (iAs). However, there is little data on the capacity of human liver to methylate iAs. This work examined the metabolism of arsenite (iAs(III)), arsenate (iAs(V)), methylarsine oxide (MAs(III)O), methylarsonic acid (MAs(V)), dimethylarsinous acid (DMAs(III)), and dimethylarsinic acid (DMAs(V)) in primary cultures of normal human hepatocytes. Primary rat hepatocytes were used as methylating controls. iAs(III) and MAs(III)O were metabolized more extensively than iAs(V) and MAs(V) by either cell type. Neither human nor rat hepatocytes metabolized DMAs(III) or DMAs(V). Methylation of iAs(III) by human hepatocytes yielded methylarsenic (MAs) and dimethylarsenic (DMAs) species; MAs(III)O was converted to DMAs. The total methylation yield (MAs and DMAs) increased over the range of 0.1 to 4 microM iAs(III). However, DMAs production was inhibited by iAs(III) in a concentration-dependent manner, and the DMAs/MAs ratio decreased. iAs(III) (10 and 20 microM) inhibited both methylation reactions. Inhibition of DMAs synthesis resulted in accumulation of iAs and MAs in human hepatocytes, suggesting that dimethylation is required for iAs clearance from cells. Methylation capacities of human hepatocytes obtained from four donors ranged from 3.1 to 35.7 pmol of iAs(III) per 10(6) cells per hour and were substantially lower than the methylation capacity of rat hepatocytes (387 pmol of iAs(III) per 10(6) cells per hour). The maximal methylation rates for either rat or human hepatocytes were attained between 0.4 and 4 microM iAs(III). In summary, (i) human hepatocytes methylate iAs, (ii) the capacities for iAs methylation vary among individuals and are saturable, and (iii) moderate concentrations of iAs inhibit DMAs synthesis, resulting in an accumulation of iAs and MAs in cells.

Biliary excretion in primary rat hepatocytes cultured in a collagen-sandwich configuration

The objective of the present investigation was to examine the functional reestablishment of polarity in freshly isolated hepatocytes cultured between 2 layers of gelled collagen (sandwich configuration). Immunoblot analysis demonstrated that the canalicular multispecific organic anion transport protein (multidrug resistance-associated protein, Mrp2) was partially maintained in day 5 hepatocytes cultured in a sandwich configuration. Fluorescein-labeled taurocholate and carboxydichlorofluorescein were excreted into and concentrated in the bile canalicular lumen of day 5 sandwich-cultured hepatocytes, resulting in formation of fluorescent networks in standard buffer (intact bile canaliculi). Confocal microscopy studies demonstrated that 1) carboxydichlorofluorescein that had concentrated in the canalicular lumen was released into the incubation buffer in the presence of Ca(2+)-free buffer (disrupted bile canaliculi), and 2) rhodamine-dextran, an extracellular space marker, was only able to diffuse into the canalicular lumen in the presence of Ca(2+)-free buffer. The cumulative uptake of [(3)H]taurocholate in day 5 sandwich-cultured hepatocytes was significantly higher in standard buffer compared with Ca(2+)-free buffer, due to accumulation of taurocholate in canalicular spaces. When [(3)H]taurocholate was preloaded in the day 5 sandwich-cultured hepatocytes, taurocholate efflux was greater in Ca(2+)-free compared with standard buffer. The biliary excretion index of taurocholate, equivalent to the percentage of retained taurocholate in the canalicular networks, increased from approximately 8% at day 0 to approximately 60% at day 5 in sandwich-cultured hepatocytes. In summary, hepatocytes cultured in a collagen-sandwich configuration for up to 5 days establish intact canalicular networks, maintain Mrp2, reestablish polarized excretion of organic anions and bile acids, and represent a useful in vitro model system to investigate the hepatobiliary disposition of substrates.

Correlation of biliary excretion in sandwich-cultured rat hepatocytes and in vivo in rats

The relationship between biliary excretion in sandwich-cultured rat hepatocytes and in vivo in rats was examined. The biliary excretion of seven model substrates in 96-h sandwich-cultured rat hepatocytes was determined by differential cumulative uptake of substrate in the monolayers preincubated in standard buffer (intact bile canaliculi) and Ca2+-free buffer (disrupted bile canaliculi). Biliary excretion in vivo was quantitated in bile duct-cannulated rats. The biliary excretion index of model substrates, equivalent to the percentage of retained substrate in the canalicular networks, was consistent with the percentage of the dose excreted in bile from in vivo experiments. The in vitro biliary clearance of inulin, salicylate, methotrexate, [D-pen2,5]enkephalin, and taurocholate, calculated as the ratio of the amount excreted into the bile canalicular networks and the area under the incubation medium concentration-time profile ( approximately 0, approximately 0, 4.1 +/- 1.0, 12.6 +/- 2.2, and 56. 2 +/- 6.0 ml/min/kg, respectively), correlated with their intrinsic in vivo biliary clearance (0.04, 0, 17.3, 34.4, and 116.9 ml/min/kg, respectively; r2 = 0.99). The model compound 264W94 was not excreted in bile either in vivo or in vitro. The glucuronide conjugate of 2169W94, the O-demethylated metabolite of 264W94, was excreted into bile in vitro when 2169W94, but not 264W94, was incubated with the monolayers; 2169W94 glucuronide undergoes extensive biliary excretion after administration of 264W94 or 2169W94 in vivo. Biliary excretion in long-term sandwich-cultured rat hepatocytes correlates with in vivo biliary excretion. The study of biliary excretion of metabolites in the hepatocyte monolayers requires consideration of the status of metabolic activities.

Use of Ca2+ modulation to evaluate biliary excretion in sandwich-cultured rat hepatocytes

Previous work in our laboratory has indicated that biliary excretion of a substrate in sandwich-cultured hepatocytes can be quantitated by measurement of substrate accumulation in the presence and absence of extracellular Ca2+. The present study was designed to examine the effects of Ca2+ on taurocholate accumulation and tight junction integrity in cultured hepatocytes. Kinetic modeling was used to characterize taurocholate disposition in the hepatocyte monolayers in the presence and absence of extracellular Ca2+. The accumulation of taurocholate in freshly isolated hepatocytes, which lack an intact canalicular network, was the same in the presence and absence of extracellular Ca2+. Electron microscopy studies showed that Ca2+ depletion increased the permeability of the tight junctions to ruthenium red, demonstrating that tight junctions were the major diffusional barrier between the canalicular lumen and the extracellular space. Cell morphology and substrate accumulation studies in the monolayers indicated that Ca2+ depletion disrupted the tight junctions in 1 to 2 min. The integrity of the disrupted tight junctions was not re-established completely after reincubation in the presence of Ca2+ for 1 h. The accumulation of taurocholate was described best by a two-compartment model (cytosol and bile) with Michaelis-Menten kinetics for both uptake and biliary excretion. In summary, Ca2+ depletion does not alter hepatocyte transport properties of taurocholate. Ca2+ modulation may be a useful approach to study biliary excretion of substrates in sandwich-cultured hepatocytes.

Partial maintenance of taurocholate uptake by adult rat hepatocytes cultured in a collagen sandwich configuration

PURPOSE

This study was designed to characterize taurocholate uptake properties in primary cultures of rat hepatocytes maintained under different matrix conditions.

METHODS

Hepatocytes isolated from male Wistar rats (230-280 g) were cultured on a simple collagen film, on a substratum of gelled collagen or between two layers of gelled collagen (sandwich configuration). Hepatocyte morphology, taurocholate uptake properties, and expression of the sinusoidal transport protein. Na+/taurocholate-cotransporting polypeptide (Ntcp) were examined in these cultures at day 0 and day 5.

RESULTS

By day 5, monolayer integrity had deteriorated in simple collagen cultures. In contrast, cell morphology was preserved in hepatocytes maintained in a sandwich configuration. At day 5, taurocholate accumulation at 5 min in hepatocytes cultured on a simple collagen film, on a substratum of gelled collagen, and in a sandwich configuration was approximately 13%, 20% and 35% of day-0 levels, respectively, and occurred predominately by a Na+-dependent mechanism. The initial taurocholate uptake rate vs. concentration (1-200 microM) profile was best described by a combined Michaelis-Menten and first-order function. In all cases, the estimated apparent Km values were comparable for day-0 and day-5 hepatocytes (3241 microM). In contrast, the Vmax values of hepatocytes cultured on a simple collagen film, on gelled collagen and in a sandwich configuration were approximately 5, 6 and 14% of the values at day 0, respectively; values for the first-order rate constant were 5-, 3- and 2-fold lower, respectively. Immunoblot analysis indicated that at day 5 Ntcp expression in hepatocytes cultured in a sandwich configuration was greater than in hepatocytes cultured on a simple collagen film.

CONCLUSIONS

A collagen sandwich configuration reestablishes normal morphology and partially restores bile acid uptake properties in primary cultures of rat hepatocytes.

Sulfotransferase gene expression in primary cultures of rat hepatocytes

Hepatocyte cultures have been used in pharmacotoxicological studies, and sulfotransferases (ST) are important drug-metabolizing enzymes in liver. The expression of sulfotransferases in hepatocyte cultures has not been examined systematically. In the present study, the mRNA levels of different sulfotransferases in male and female rat hepatocytes were examined by northern-blot analyses. Various culture conditions such as different matrices (collagen, matrigel, collagen sandwich, or co-culture with epithelial cells), medium (Way-mouth's MB 752/1 and Modified Chee's Medium) and glucocorticoid supplementation (dexamethasone, 0.1 microM) were compared. Phenol ST (ST1A1) mRNA levels decreased to about 50% of initial mRNA levels within 10 hr of culture. At 96 hr, ST1A1 mRNA levels were approximately 20% of initial values when cultured on collagen, matrigel or co-culture. The two media did not differ in ability to maintain ST1A1 mRNA levels in the absence of dexamethasone (DEX); however, DEX addition to either medium resulted in ST1A1 mRNA levels greater than 100% of the initial mRNA levels at 96 hr, with the greatest increase observed using the matrigel substratum and Chee's medium. In the absence of DEX, the mRNA levels of N-hydroxy-2-acetylaminoflurene sulfortransferase (ST1C1), estrogen sulfotransferase (ST1E2) and hydroxysteroid sulfotransferase (ST-20/21, ST-40/41, ST-60) fell to approximately 20% of their initial levels within 24 hr, and to less than 5% at 96 hr. The loss of expression of these sulfotransferases was observed with all culture conditions. Addition of DEX to the media resulted in ST-40/41 and ST-60 mRNA expression at 20 and 35% of their initial values, respectively, in cultures maintained on matrigel and Chee's medium at 96 hr. These data suggest that sulfotransferases lose their constitutive expression in hepatocyte culture, but retain their inducibility.

Assessment of a model for measuring drug diffusion through implant-generated fibrous capsule membranes

Fibrous tissue, which encapsulates subcutaneously implanted silastic, vinyl, polyurethane and Teflon discs in rats, has been isolated, characterized and tested for drug permeability in order to develop an in vitro model for determining the effect of this tissue on drug disposition from implant sites. With all materials, capsule tissue thickness and collagen content (approximately 59%) was consistent from 2 to 4 months after implantation. Silastic implants afforded the most consistent and usable tissue in terms of thickness and lack of vascularity, and these capsule membranes were used for determining the transport of three model compounds in an in vitro diffusion cell model. The rank ordering of permeability through these membranes was estrone (60.2 x 10(-6) cm s-1) > 3-O-methylglucose (18.7 x 10(-6) cm s-1) > dextran of molecular weight 70 000 (5.6 x 10(-6) cm s-1), which is consistent with expectations based on the molecular weights and partitioning behaviour of the model compounds. The results of these studies indicate that implant-generated encapsulating membranes can be successfully isolated and employed to study drug diffusion in an in vitro model, providing a direct assessment of the barrier properties of encapsulating membranes.

Formation of extensive canalicular networks by rat hepatocytes cultured in collagen-sandwich configuration

Rat primary hepatocytes were cultured under different extracellular matrix configurations and evaluated for the acquisition and maintenance of structural and functional cell polarity. De novo repolarization of the plasma membrane was variable in rate and extent in hepatocyte cultures maintained on a conventional single layer of either gelled or ungelled collagen. However, cultures maintained in a collagen-sandwich configuration initiated uniform formation of a contiguous anastomosing network of bile canaliculi throughout the entire culture. Localization of apical membrane markers demonstrated normal distribution at the canalicular membrane. A marked rearrangement of the intracellular microfilaments to the cell periphery was observed and coincided with the development of the bile canaliculi. Acquisition of normal bile canalicular function and integrity was observed within 3-4 days postoverlay as indicated by the concentration and retention of carboxyfluorescein within the canalicular network. These results demonstrate that cultures of hepatocytes maintained in a sandwich configuration may serve as a more reliable and representative model in which to study the physiology of hepatic function as well as the morphogenesis of polarized membrane domains in vitro.

In vitro and in vivo analysis of the mechanism of absorption enhancement by palmitoylcarnitine

Long-chain acylcarnitines (12-18 carbon fatty acid esters) dramatically enhance the absorption of hydrophilic drugs across intestinal mucosa without altering the morphology of the epithelium. The mechanism underlying these effects was studied using the colon carcinoma cell line Caco-2. Caco-2 monolayers treated with 0.2 mM palmitoylcarnitine (PCC) show dramatic increases in the transport of hydrophilic markers. This enhanced transport coincides with a rapid drop in transepithelial electrical resistance (TER). The drop in TER is initiated within the first minute after PCC addition and continues for approximately 20 min to a 70 to 85% drop of the initial TER values. This effect is reversible after removing the PCC and does not appear to involve lysis of the apical membrane. Instead PCC's effect appears to be due to loosening of tight junctions as indicated by the accumulation of fluorescent dextrans and the electron dense marker lanthanum nitrate in paracellular spaces. Moreover transmission electron microscopy and freeze fracture electron microscopy indicate that PCC produces significant structural alterations to tight junctions. In contrast to many other tight junction disrupting agents, PCC effects appear to be Ca(++)-independent and PCC does not induce significant disruption of actin filament distribution in Caco-2 cells.

Enhanced bioavailability of cefoxitin using palmitoylcarnitine. II. Use of directly compressed tablet formulations in the rat and dog

The performance of tablets containing the absorption enhancer palmitoylcarnitine chloride (PCC) and the antibiotic cefoxitin (CEF) was determined by direct placement of tablets in the rat stomach, small intestine, and colon. While the bioavailability (F) of tablets containing 12 mg CEF without PCC ranged from 0.6 to 3.9%, the addition of 24 mg PCC resulted in an enhanced CEF bioavailability in the rat colon (mean +/- SD: F = 57 +/- 19%) and rat jejunum (F = 71 +/- 16%) but not in the rat stomach. Following oral administration to dogs, tablets of 200 mg CEF without or with 600 mg PCC resulted in the same low bioavailabilities (7.0 +/- 10.3 and 7.0 +/- 3.6%, respectively). However, when these tablets were enteric coated, PCC improved CEF bioavailability from 2.44 +/- 1.84 to 29.0 +/- 13.4%. Therefore, the use of enteric-coated direct compressed tablets containing PCC and direct compression excipients improved the peroral bioavailability of a poorly absorbed compound.

Simultaneous in vitro measurement of intestinal tissue permeability and transepithelial electrical resistance (TEER) using Sweetana-Grass diffusion cells

A simple modification of the commercially available Sweetana-Grass (S-G) side-by-side diffusion cells, allowing the simultaneous measurement of tissue permeability and transepithelial electrical resistance (TEER), has been described and validated for rat excised, muscle-free intestinal tissue. The TEER-lowering effects of a series of acylcarnitines were shown to be correlated with previously reported in vitro (i.e., membrane perturbation) and in vivo (i.e., absorption enhancement) activity. The TEER-lowering effect of palmitoyl carnitine chloride (PCC) was also shown to be reversible. The effects of PCC on TEER and the permeability of poorly absorbed compounds (cefoxitin and lucifer yellow) were simultaneously determined. Compared to controls (mannitol-treated), PCC immediately produced a rapid drop in colon TEER. By 5 min post-PCC addition, colon TEER was 50% of control; by 10 min post-PCC addition, colon TEER was 17% of control. After a lag of about 5-10 min post-PCC addition, the cefoxitin or lucifer yellow permeability coefficient increased more than 20-fold. The modified S-G cells provide a simple and reproducible method whereby flux and TEER can be simultaneously determined, providing a valuable link between the effect of absorption enhancers on TEER measurements and the increased permeability of poorly absorbed compounds.

Enhanced bioavailability of cefoxitin using palmitoyl L-carnitine. I. Enhancer activity in different intestinal regions

The conditions under which the absorption enhancer palmitoyl L-carnitine chloride (PCC) improved the bioavailability of the poorly absorbed antibiotic cefoxitin throughout the rat intestine has been studied. Cefoxitin alone was appreciably absorbed only in the duodenum (31% vs less than 7% elsewhere). PCC solutions (3 mg/rat, pH 4.0) enhanced cefoxitin bioavailability (F) by 0-, 22-, 16-, and greater than 32-fold in the duodenum, jejunum, ileum, and colon regions, respectively. The inability of PCC to improve F in the duodenum could not likely be attributed to enzymatic degradation of the enhancer, since coadministration with protease and esterase inhibitors produced similar results (F = 30%). Coadministration of PCC solution with cefoxitin in the unligated or ligated colon, increased F to 33 and 76%, respectively. Qualitatively similar results were seen with PCC suspensions (3 mg/rat, pH 6.0). Maintaining a high concentration of cefoxitin and PCC in a restricted region (i.e., by ligating a 2- to 3-cm section of the colon) afforded a two- to threefold advantage over an unligated colon section. The difference in cefoxitin bioavailability between ligated and unligated colon was probably due to sample spreading and subsequent/simultaneous dilution.

Relationship between drug absorption enhancing activity and membrane perturbing effects of acylcarnitines

Acylcarnitines with chain lengths of 2 to 18 carbon atoms were tested for their effects on rat intestinal brush border membrane order (S) by fluorescence polarization of 1,6-diphenyl-1,3,5-hexatriene (DPH). These results were compared to the previously reported effectiveness of the acylcarnitines as absorption enhancers of the poorly absorbed antibiotic cefoxitin. Acylcarnitines with fatty acids less than 12 carbon units in length were ineffective in increasing drug absorption and perturbing brush border membrane order. Long-chain acylcarnitines (12-18 carbons) significantly increased the bioavailability of cefoxitin and decreased the lipid order of brush border membranes. The results suggest that, in order to promote drug absorption, the acylcarnitines must surpass a critical chain length (10 carbon units) to partition effectively into the membrane and, in addition, must perturb the lipid order beyond a threshold value (15-20%). Membrane perturbing capacity may serve as an indicator of the absorption enhancing potential of other aliphatic-type compounds.

Development of microtubule capping structures in ciliated epithelial cells

Although capping structures are present at the tips of microtubules in both growing cilia and mature cilia, previous work has not determined the time of cap formation. The results reported here reveal that the large caps of mature palate cilia appear in cilia with lengths as short as 1.75 micron. In the growing palate cilium, a disk-shaped plate is formed at the tip during the first micron of growth. As the cilium elongates to 1.5-2.0 microns, a small plate forms underneath the disk-shaped plate that gives an asymmetrical appearance to the whole cap structure. The structure of the cap is complete in cilia longer than 2.0 microns. The hair-like structures that form the extraciliary crown appear on the membrane at the ciliary tip at the same time as the mature cap is forming. The formation of a cap structure is discussed in relation to microtubule assembly during ciliogenesis.

Development and ciliation of the palate in two frogs, Bombina and Xenopus; a comparative study

The morphological changes that occur during metamorphosis in the palates of two types of anuran larvae (a discoglossid, Bombina orientalis, and a pipid, Xenopus laevis) are compared. In B. orientalis the structural changes are accompanied by the ciliation of the palate epithelium. Ciliation begins in the anterior region of the palate and continues in a posterior direction throughout metamorphosis. By contrast, the palate of X. laevis never becomes ciliated during its development. Instead, two ciliated grooves develop between the choanae (nasal openings) and the esophageal opening. The grooves transport mucus and trapped objects out of the internal nares and toward the esophagus. These grooves are compared to similar structures on the palate of adult B. orientalis. The timing and pattern of ciliogenesis during metamorphosis in each of these anurans is also described relative to well-established staging series for external frog development. We show that the onset and location of ciliogenesis are consistent and predictable in these anurans and, therefore, make the frog palate an excellent system for the study of ciliogenesis.

Asymmetrical microtubule capping structures in frog palate cilia

Capping structures at the distal tips of frog palate cilia are attached to the A- and central pair microtubules by electron-dense plug structures similar to those found in protozoan cilia and flagella and in epithelial cilia from invertebrates and vertebrates (W.L. Dentler, 1980, J. Cell Sci. 42, 207-220; W.L. Dentler and E.L. LeCluyse, 1982, Cell Motil. 2, 549-573). The caps in frog palate cilia are composed of a proximal shelf, to which doublets Nos. 1-3, 8, 9 and the central microtubules are attached and a larger distal cap to which doublets Nos. 4-7 are bound. The smaller proximal shelf is positioned to one side of the cilium and gives the cap an asymmetrical appearance. Striated ciliary rootlets attached to the basal bodies are also described. The smaller cap is placed on the same side of all cilia on the palate relative to both the direction of the effective stroke and the position of the rootlets. These results confirm that capping structures are common to most, if not all, cilia and provide direct evidence that they are precisely positioned on specific microtubules.

Microtubule capping structures at the tips of tracheal cilia: evidence for their firm attachment during ciliary bend formation and the restriction of microtubule sliding

The distal tips of the central pair and A-microtubules are capped in mammalian and avian tracheal cilia. The capping structures are similar to those found in protozoan cilia and flagella [Dentler, 1981], and consist of a central microtubule cap that links the central microtubules to the membrane or to the ciliary crown and A-microtubule plugs that insert into the lumen of each of the A-microtubules. In contrast to the protozoan capping structures, each of the A-microtubule plugs is bound to the central microtubule cap by distal filaments. The ends of the central and outer doublet microtubules are tightly bound to the cap in both intact and in demembranated and reactivated tracheal cilia. Analysis of the displacement of the microtubule tips in cilia fixed at various bend angles revealed that the displacements of A-microtubules are only partially in agreement with those predicted by the sliding filament model [Satir, 1968]. These results are discussed with respect to the regulation of microtubule sliding in capped cilia and the role of the microtubule capping structures in microtubule assembly.

Microtubule-associated proteins (MAPs) and the organization of actin filaments in vitro

When purified muscle actin was mixed with microtubule-associated proteins (MAPs) prepared from brain microtubules assembled in vitro, actin filaments were organized into discrete bundles, 26 nm in diameter. MAP-2 was the principal protein necessary for the formation of the bundles. Analysis of MAP-actin bundle formation by sedimentation and electrophoresis revealed the bundles to be composed of approximately 20% MAP-2 and 80% actin by weight. Transverse striations were observed to occur at 28-nm intervals along negatively stained MAP-actin bundles, and short projections, approximately 12 nm long and spaced at 28-nm intervals, were resolved by high-resolution metal shadowing. The formation of MAP-actin bundles was inhibited by millimolar concentrations of ATP, AMP-PCP (beta, gamma-methylene-adenosine triphosphate), and pyrophosphate but not by AMP, ADP, or GTP. The addition of ATP to a solution containing MAP-actin bundles resulted in the dissociation of the bundles into individual actin filaments; discrete particles, presumably MAP-2, were periodically attached along the splayed filaments. These results demonstrate that MAPs can bind to actin filaments and can induce the reversible formation of actin filament bundles in vitro.