hexamethylene bisacetamide protects peritoneal mesothelial cells from glucose

Kidney International, Vol. 60 (2001), pp. 996–1008

Hexamethylene bisacetamide protects peritoneal mesothelialcells from glucose

TAKAHIKO OGAWA, TOMONORI HAYASHI, NORIAKI YORIOKA, SEISHI KYOIZUMI,and JAMES E. TROSKO

Nephrology and Dialysis Division, Hiroshima Prefectural Hospital, Department of Radiobiology, Radiation Effects ResearchFoundation, and Second Department of Internal Medicine, Hiroshima University School of Medicine, Hiroshima, Japan;and National Food Safety Toxicology Center, Department of Pediatrics/Human Development, Michigan State University,East Lansing, Michigan, USA

significant change of Cx43 phosphorylation in the presence ofHexamethylene bisacetamide protects peritoneal mesothelialthese inhibitors.cells from glucose.

Conclusions. High glucose down-regulates GJIC in humanBackground. Peritoneal dialysis causes damage to perito-peritoneal mesothelial cells. It also decreases the levels of bothneal mesothelial cells primarily because dialysis fluids have aCx43 mRNA and Cx43 protein, with the latter becoming hypo-high glucose concentration. This study examined the abnormal-phosphorylated. HMBA appears to reverse all of these changes.ities of gap junctional intercellular communication (GJIC) inThese results are consistent with our hypothesis that HMBAhuman peritoneal mesothelial cells (HPMCs) exposed to rela-protects HPMCs from the adverse effects of high glucose bytively high levels of glucose. Also, ability of hexamethylene bis-reversing various processes that would otherwise lead to harm-acetamide (HMBA) to up-regulate GJIC in HPMCs exposedful loss of GJIC.to high levels of glucose was measured.

Methods. An assay that monitors the recovery of fluorescenceafter photobleaching was used to measure GJIC in primarycultured HPMCs. The cells were exposed to a low (10 mmol/L) Peritoneal dialysis is a popular alternative to hemodialy-or high (50 or 90 mmol/L) glucose level for a total of six days,

sis. It employs the human peritoneum as a living dialysisand some cells were also incubated with or without HMBA (1membrane, the permeability of which can be modified byor 6 mmol/L) from day 4. The effects of incubation in these

various environments on expression of the connexin 43 (Cx43) various biological and physical factors [1–3]. Peritonealgene were investigated by the reverse transcription-polymerase mesothelial cells are thought to have a vital role in main-chain reaction (to detect Cx43 mRNA) or by immunofluores- taining the integrity of the peritoneum, and intercellularcence and Western blotting (to detect Cx43 protein). To evalu-

adhesive interactions also contribute to peritoneal struc-ate the influence of protein kinase C (PKC) or mitogen-acti-tural integrity [3–5]. However, prolonged exposure to dial-vated protein kinase (MAPK) on GJIC, specific inhibitors were

added to cultures in a high glucose medium. ysis fluid may damage peritoneal mesothelial cells [2,Results. Gap junctional intercellular communication was in- 6–10]. Not only the direct toxic effects of dialysis fluid

hibited in a concentration- and time-dependent manner when (related to factors such as a low pH, a high glucose concen-cells were exposed to high glucose. The addition of 6 mmol/Ltration, and hyperosmolarity), but also indirect effects re-HMBA to cultures significantly enhanced GJIC despite thelated to disturbance of intercellular humoral interactionspresence of a high glucose concentration. High glucose also

down-regulated Cx43 mRNA and protein expression, with the within the peritoneal cavity, may injure the peritoneumdose-dependent decrease of Cx43 protein at gap junctions par- [7, 8, 11, 12] and lead to the loss of peritoneal mesothelialalleled by a decrease in the phosphorylation of this protein. cells [2, 6]. It has been shown that creating a wound inAs expected, treatment of cells with 6 mmol/L HMBA in-

the peritoneum induces the migration and proliferationcreased both Cx43 mRNA and protein levels despite exposureof mesothelial cells into the damaged area [13]. Intercel-to high glucose. The addition of PKC or MAPK inhibitors to

high glucose cultures did not restore GJIC, and there was no lular communication may be important in this process,just as it is for other kinds of cells [14]. Sammak et alhave suggested that gap junction-mediated diffusion ofKey words: gap junctional intracellular communication, connexin 43,intracellular messengers between injured cells and nor-protein kinase inhibitors, dialysate, renoprotection, intracellular mes-

sengers, cell signaling. mal cells may be responsible for the promotion of woundhealing [15]. Although little is known about whetherReceived for publication May 26, 2000the loss of gap junctional intercellular communicationand in revised form March 19, 2001

Accepted for publication April 2, 2001 (GJIC) is likely to have specific adverse effects on perito-neal mesothelial cells, numerous reports show that alter- 2001 by the International Society of Nephrology

996

Ogawa et al: Glucose and HMBA in mesothelial cells 997

ation of GJIC is associated with human disease [reviewed regulation of GJIC [23, 24, 31–33], it is clearly importantto determine whether high glucose interferes with GJICin 16]. Therefore, it seems important to study the effects

of peritoneal dialysis fluids on GJIC in peritoneal meso- in HPMCs via its ability to activate one or more of thepreviously mentioned kinase pathways.thelial cells, as suggested by Nagy [3].

Gap junctions are clusters of transmembrane channels The present study addressed the following issues. First,does glucose inhibit GJIC in human peritoneal mesothe-composed of proteins called connexins that permit the

direct exchange of second messengers between adjacent lial cells? Second, if it does inhibit GJIC, what is themechanism involved? Third, is an intracellular signalcells [17, 18]. This gap junctional exchange has been shown

to be an important method by which cells communicate transduction system involving PKC and/or MAPK in-volved? Fourth, can HMBA restore GJIC in high glu-and regulate many of their vital functions, including re-

generation and proliferation [17]. Studies have been per- cose-inhibited human peritoneal mesothelial cells? Toinvestigate these points, we examined GJIC in HPMCsformed to assess the effect of various stimuli on GJIC, and

many chemicals have been shown to inhibit this process cultured with various concentrations of glucose and/orHMBA and a range of putative enzyme inhibitors.[19–21], while others appear to up-regulate it [22].

Recent studies on diabetic vasculopathy have shownthat high glucose levels reduce gap junction permeability

METHODSin cultured vascular endothelial cells and smooth muscle

Isolation and culture of human peritonealcells [23, 24]. Inhibition of GJIC may lead to dysregula-mesothelial cellstion of cell regeneration and proliferation [16]. In the

case of human peritoneal mesothelial cells (HPMCs), Human peritoneal mesothelial cells were harvestedfrom the omental tissues of five consenting patients whohowever, the relationship between glucose and GJIC

remains largely unknown. Some reports have suggested underwent elective abdominal surgery. As described pre-viously [28, 34], the cells were isolated and cultured inthat relatively low glucose concentrations corresponding

to the level of hyperglycemia commonly seen in diabetic basal medium, which was Ham’s F-12 medium (10 mmol/Lglucose; GIBCO BRL, Grand Island, NY, USA) supple-patients may be enough to produce cellular injury [25, 26]

and that higher glucose concentrations may be capable mented with penicillin (100 U/mL; GIBCO BRL), strepto-mycin (100 �g/mL; GIBCO BRL), l-glutamine (2 mmol/L;of causing marked impairment of mesothelial cell prolif-

eration [27]. To begin to understand the effects of glucose GIBCO BRL), and 10% fetal calf serum (FCS; IntergenCo., Purchase, NY, USA). All experiments were done us-on GJIC—which may influence its role in the mainte-

nance of peritoneal homeostasis and in certain cellular ing cells from primary cultures or from the second passage.repair mechanisms—much more information about the

Experimental designeffects of the very high glucose concentrations found inperitoneal dialysis fluid on GJIC in HPMCs is required. Cells were seeded at a density of 1.0 � 104/cm2 in basal

medium. After confluence was reached, the medium wasWe recently reported that HPMCs cultured in thepresence of a low (10 mmol/L) glucose concentration removed, and new conditioning medium containing glu-

cose and/or HMBA (Sigma Chemical Co., St Louis, MO,show active GJIC, which seems to be mediated by con-nexin 43 (Cx43), a primary gap junction protein that has USA) was added. Glucose concentrations in the medium

were selected to match the glucose levels in commerciallymore than one active phosphorylated form as well as asingle inactive unphosphorylated form [28]. That study available peritoneal dialysis fluids. The basal medium

was changed every three days, while the conditioningalso showed that hexamethylene bisacetamide (HMBA),a hybrid polar compound known as a potent differentia- medium was changed daily to maintain the concentra-

tions of glucose and HMBA.tion inducer, can up-regulate GJIC in HPMCs, and thatthis effect of HMBA is related to an increase in the Four experimental protocols were used. Protocol A

assessed the cytotoxicity of glucose and the changes ofactive form of Cx43 (through post-translational phos-phorylation) as well as to an increase in Cx43 mRNA GJIC during culture of HPMCs with high glucose me-

dium. Basal medium was replaced by conditioning me-expression [28]. Therefore, HMBA may have the poten-tial to exert a protective effect on the peritoneum by dium with a final glucose concentration of either 50 or

90 mmol/L, and culture was continued for six days, threesignificantly enhancing GJIC among mesothelial cells.It also has been reported that high glucose levels can days, or one day, while control cells were cultured in

basal medium for six days. To assess the influence ofactivate cellular signaling processes involving protein ki-nase C (PKC) and/or mitogen-activated protein kinase hyperosmolarity on GJIC, mannitol was used as a non-

metabolized osmotic agent. Basal medium (10 mmol/L(MAPK), including (and perhaps especially) extracellu-lar signal-regulated kinase (ERK) in many kinds of cells glucose) plus 40 mmol/L mannitol was employed to cre-

ate an osmotic load equivalent to that of 50 mmol/L[29, 30]. Given that Cx43 is a PKC/MAPK substrateand that phosphorylation of Cx43 can lead to down- glucose, while basal medium plus 80 mmol/L mannitol

Ogawa et al: Glucose and HMBA in mesothelial cells998

was employed to create an osmotic load equivalent to the present experiments, HPMCs were seeded at a den-sity of 5.0 � 103 cells per well. After confluence was90 mmol/L glucose. Protocol B assessed the effects of

HMBA on GJIC under high-glucose conditions. The reached, the cells were exposed to conditioning medium.WST-1 staining was done on day 6 in protocol A. Resultsincubation times and the concentrations of HMBA used

were based on the results of earlier studies [28, 35–38]. were expressed as the mean % absorbance � the SEMrelative to the value for cells cultured with 10 mmol/LWe previously found that incubation with 3 or 6 mmol/L

HMBA for three days and with 1 mmol/L HMBA for six glucose. To assess the toxicity of HMBA, it was addedto the basal medium at final concentrations of 1, 3, or 6days caused significant up-regulation of GJIC in HPMCs

[28]. Therefore, cells in the present study were incubated mmol/L, and WST-1 staining was done after three days.for three days with 1 and 6 mmol/L HMBA. The former

Fluorescence recovery after photobleachingconcentration was not high enough to up-regulate GJIC,assay for GJICbut the latter was sufficient to do so; the results were

compared with those obtained in the absence of HMBA. The procedure used was a modification of that origi-nally employed for measurement of GJIC by the quanti-After incubation of HPMCs for three days with 10, 50,

or 90 mmol/L glucose, the medium was replaced with tative fluorescence recovery after photobleaching (FRAP)assay [43]. Assays were performed using an ACAS Ul-HMBA-containing medium (0, 1, or 6 mmol/L), and

culture was continued for another three days. Protocol tima laser cytometer (Meridian Instruments, Inc., Okemos,MI, USA). After selectively bleaching cells with a micro-C was used to assess the effects of high glucose on GJIC

after pretreatment with HMBA. After incubation of laser beam, the rate of transfer of 5, 6-carboxyfluoresceindiacetate (Molecular Probes, Inc., Eugene, OR, USA)HPMCs with 6 mmol/L HMBA for three days, the me-

dium was replaced with medium containing both HMBA from the adjacent labeled cell back into the bleachedcell was calculated. The recovery of fluorescence wasand glucose (10, 50, or 90 mmol/L), and culture was contin-

ued for another three days. In both protocol B and proto- examined after one minute, and the recovery rate wascalculated as %/min (% � the percentage of photo-col C, all analyses were done on day six of culture. Protocol

D was used to clarify the role of the PKC/MAPK cascade bleached fluorescence). The recovery rate was correctedfor the loss of fluorescence by unbleached control cells,in GJIC. Two protein kinase inhibitors and one activator

were used. Calphostin C (50 nmol/L; Calbiochem, La and then, results were expressed as the average percent-age of recovery rate � SEM relative to that of controlJolla, CA, USA) was used as a specific inhibitor of PKC

[39]. PD98059 [2�-amino-3�-methoxyflavone] (20 �mol/L; cells incubated with 10 mmol/L glucose.Calbiochem), a specific inhibitor of MAPK/ERK kinase

Immunofluorescence microscopy(MEK), was used as a substitute for an ideal MAPKinhibitor (which would be capable of blocking all MAPK Indirect immunofluorescence microscopy was per-

formed as described previously [28, 44]. After fixing andpathways) [30, 40]. 12-O tetradecanoylphorbol-13-acetate(15 nmol/L; TPA; Sigma) is a potential activator of PKC washing, cells were incubated with a monoclonal mouse

antibody for Cx43 (diluted 1:2000; Chemicon Interna-and MAPK [29, 30]. To assess the effects of the proteinkinase inhibitors during prolonged high glucose treat- tional Inc., Temecula, CA, USA), followed by incubation

with a fluorescein isothiocyanate-conjugated secondaryment, HPMCs were incubated for three days with 10 or90 mmol/L glucose, after which the medium was replaced anti-mouse antibody (diluted 1:200; Biosource, Cama-

rillo, CA, USA). The primary antibody was raised againstwith conditioning medium containing an inhibitor (or anactivator, where appropriate), and incubation was contin- a peptide corresponding to residues 252 to 270 of the

native sequence of Cx43 [45]. This antibody showed noued for another four hours (one hour for TPA). It shouldbe noted that the incubation time and the concentrations staining after preabsorption with the corresponding syn-

thetic polypeptide (data not shown). Specific stainingof inhibitors or the activator were all decided on thebasis of previous experimental findings [24, 30, 41]. was identified by comparison with negative controls that

were processed with normal mouse IgG and phosphate-Cytotoxicity assay buffered saline (PBS). Immunofluorescence micrographs

were obtained using an epifluorescence microscopeTo assess the cytotoxicity of glucose and HMBA forHPMCs, viable cells were detected by staining with a equipped for photography and Fujichrome film (Fujifilm,

Tokyo, Japan).tetrazolium salt, 4-[3-(4-iodophenyl)-2-(4-nitrophenyl)-2H-5-tetrazolio]-1,3-benzene disulfonate (WST-1; Do-

Extraction of Cx43 protein and RNAjindo Laboratories, Kumamoto, Japan) [42]. In prelimi-nary experiments, WST-1 staining of HPMCs (shown by Cells were grown in 10 cm dishes and were prepared

as described previously [28]. In brief, after six days ofthe absorbance at 450 nm) was found to increase linearlywith the number of viable cells from 1.0 � 103 to 1.0 � incubation, the cells were washed three times with PBS.

TRIzol reagent (GIBCO BRL) was added to each dish,105 cells per well in a 96-well plate (data not shown). In


and the lysates were stored in a freezer until use. RNA transcribed. The average control value was then arbi-was obtained from the upper aqueous phase of the TRI- trarily set at one, and all readings were expressed relativezol reagent extract and was dissolved in RNase-free wa- to this in arbitrary units. A negative control reaction mix-ter, after which the RNA concentration was determined ture without cDNA and a control without reverse tran-by spectrophotometry at 260 nm. The reverse transcrip- scription (RT�) were included in all series of reactions.tion-polymerase chain reaction (RT-PCR) was performed

Statistical analysisas described later in this article. The protein pellet ex-tracted from the lower phenol-chloroform phase was Data were analyzed using Statview II software (Applewashed with 99.5% ethanol, centrifuged, and air dried Macintosh Inc., Cupertino, CA, USA). The two-tailed,for use in the following assays. unpaired t test was used to compare the glucose/HMBA

or glucose/protein kinase inhibitor cultures with controlWestern blottingcultures, and differences were considered significant at

Proteins were dissolved in 0.2% sodium dodecyl sul- P � 0.05. Results are expressed as the mean � SEM forfate containing protease inhibitors and were sonicated experiments performed with HPMC isolated from fiveas described previously [28]. Protein concentrations were

different omental specimens, unless otherwise stated.determined using the DC protein assay (Bio-Rad Corp.,Richmond, CA, USA) after diluting the samples 1:5 withH2O. Twenty micrograms of protein per sample were RESULTSdissolved in Laemmli sample buffer, separated on a Cytotoxicity of glucose and HMBA12.5% polyacrylamide gel, and transferred to a polyvinil-

Phase-contrast micrographs of control mesothelialidene difluoride membrane (Bio-Rad). Cx43 was detectedcells revealed uniform monolayers of polygonal cells thatby incubation with an anti-Cx43 monoclonal antibodyclearly exhibited contact inhibition. Neither glucose (50(diluted 1:2000; Chemicon International Inc.), followedor 90 mmol/L) nor equiosmolar mannitol induced sub-by incubation with a horseradish peroxidase-conjugatedstantial morphological changes in the cells, and theresecondary antibody (diluted 1:2000, Amersham Co., Ar-was no evidence of severe cytotoxicity after incubationlington Heights, IL, USA) and an enhanced chemilumi-for six days. Table 1 shows the results of the WST-1nescence detection reagent (Renaissance Western Blotassay of cell viability. Irrespective of the incubation timeChemiluminescence Reagent; NEN Life Science Prod-

ucts, Inc., Boston, MA, USA). The intensity of the Cx43 and/or the glucose or mannitol concentration, there wereprotein bands was normalized for that of �-tubulin. The no significant changes of viability when compared withaverage control value was assigned an arbitrary value of cells incubated in 10 mmol/L glucose for the same time.one unit, and relative band intensities were standardized There was also no significant effect on viability whento this arbitrary unit. cells were incubated with HMBA for three days (Table

2). To exclude any effect of FCS on GJIC, medium con-Semiquantitative reverse transcription-polymerase taining 0.1% FCS instead of 10% FCS was also employedchain reaction assay

in preliminary experiments, but there were no significantReverse transcription (RT) and the PCR were per- differences of GJIC between cells cultured in the basal

formed as described previously [28, 46]. The oligonucleo- medium and in this medium. However, culture of cellstide primers used were a 5�-Cx43 primer for exon one for more than four days with 0.1% FCS in the presence(5�-GCGTGAGGAAAGTACCAAAC-3�) and a 3�-Cx43 of high glucose led to a decrease in viability, as assessedprimer for exon two (5�-CCTTCTTGAGTTCCAAC by the WST-1 assay (data not shown). Thus, the absenceGGG-3�). PCR products were subjected to electrophore- of FCS might modify the absolute values obtained, butsis on 8% acrylamide gel and were visualized by staining

not the nature of the changes in fluorescence recovery.with ethidium bromide, after which the relative intensity

All subsequent experiments were performed using me-of each band was determined using image analysis soft-dium containing 10% FCS.ware (NIH Image). To confirm the reproducibility of the

results, PCR and electrophoresis were repeated three GJIC after exposure to high glucose and HMBAtimes for each sample. Simultaneous amplification of

Figure 1 shows two series of typical digitized imagescDNA for the housekeeping gene glyceraldehyde-3-phos-obtained in the FRAP assay. After bleaching, sequentialphate dehydrogenase (GAPDH) always was done in thescans were done to detect the recovery of fluorescencesame tubes (using the primers 5�-CCACCCATGGCAAin bleached cells (Fig. 1). Recovery of fluorescence afterATTCCATGGCA-3� and 5�-TCTAGACGCCAGGTCphotobleaching was considerably less rapid when cellsAGGTCCACC-3�), and all results were expressed as thewere cultured with 90 mmol/L glucose (Fig. 1B) than whenratio of Cx43 cDNA to GAPDH cDNA in order to

control for variations in the amount of RNA initially cells were cultured with 10 mmol/L glucose (Fig. 1A).


Table 1. Toxicity of glucose and mannitol in the WST-1 assay

Incubation time days

Culture conditions 0 1 3 6

Glucose10 mmol/L 100.0�7.1 100.0�7.1 100.0�7.6 100.0�6.550 mmol/L 88.8�5.2 93.8�5.8 116.3�2.490 mmol/L 111.4�8.0 87.0�5.3 98.7�3.6

10 mmol/L Glucose40 mmol/L Mannitol 81.0�8.2 99.5�5.8 107.6�12.780 mmol/L Mannitol 87.4�6.0 96.4�1.9 101.0�13.8

Values are the mean � SEM of the % OD at 450 nm calculated relative to that of cells cultured with 10 mmol/L glucose (100%) for HPMCs isolated from fivedifferent donors. The final glucose or mannitol concentrations in the culture medium are shown.

Table 2. Toxicity of HMBA in the WST-1 assay high (50 or 90 mmol/L; Fig. 2B). Incubation of HMBA-pretreated cells with 50 or 90 mmol/L glucose for anHMBA mmol/Ladditional three days caused the dye transfer rate to fall0 1 3 6significantly below that observed after supplementary% OD at 450 nm 100�2.9 101.5�11.0 92.6�30.4 94.9�4.4incubation with 10 mmol/L glucose (Fig. 2C).

Values represent the mean � SEM from five separate experiments performedwith HPMCs from five different donors. % OD was calculated relative to no

Effects of high glucose and HMBA onHMBA treatment (100%).

Cx43 immunofluorescence

The localization of Cx43 protein was examined byindirect immunofluorescence cytochemistry. Figure 3In protocol A, glucose decreased the rate at which theshows immunostaining of cells for Cx43 after six days ofdye was transferred to photobleached cells by GJIC in aincubation in medium containing glucose with or withouttime- and concentration-dependent manner. IncubationHMBA for the final three days (protocol B). The nega-with 50 mmol/L glucose for three and six days, as welltive control, in which normal mouse IgG was substitutedas with 90 mmol/L glucose for one, three, and six days,for the primary antibody, showed no staining. Controlalso caused a significant decrease in the rate of dye trans-cells incubated with 10 mmol/L glucose showed severalfer by GJIC (Fig. 2A). At the same glucose concentra-bright spots indicating Cx43 labeling at the points oftion, cells incubated for six days showed a lower rate ofintercellular contact. Cells cultured with 50 mmol/L glu-dye transfer than cells cultured for one or three days.cose still displayed a few spots at the cell borders, butIn contrast, the dye transfer rate was not significantlyno labeling was seen after cells were cultured with 90decreased following incubation with 40 mmol/L manni-mmol/L glucose. Incubation with HMBA led to an in-tol. Although 80 mmol/L mannitol decreased the dyecrease in the number and size of the labeled regions,transfer rate after three or six days of incubation, itswith the cells displaying linear or dotted labeling alongeffect was weaker than that of 90 mmol/L glucose (Fig.the membrane. At the higher glucose concentration, en-2A). Incubation of glucose-treated cells with 6 mmol/Lhancement of labeling was observed with 6 mmol/LHMBA for three days caused the dye transfer rate toHMBA but not with 1 mmol/L HMBA.increase significantly compared with that observed in the

absence of HMBA (Fig. 2B). Even 1 mmol/L HMBAEffects of high glucose and HMBA on Cx43 proteincaused a dose-dependent increase of GJIC-mediated dye

Western blotting was used to determine whether GJICtransfer when the glucose concentration was low (10activity was related to the total Cx43 protein level and/ormmol/L), whereas 6 mmol/L HMBA was necessary to

increase dye transfer when the glucose concentration was to the extent of Cx43 phosphorylation. Three forms of

�

Fig. 1. Fluorescence recovery after photobleaching (FRAP) assay. Typical digitized fluorescence images and plots of fluorescence recovery afterphotobleaching are shown. After exposure to 10 (A) or 90 mmol/L (B) glucose for six days, cells were labeled with 5, 6-carboxyfluorescein diacetate.Suitable fields of cells were identified using a �40 objective lens. These fields contained numerous cells that were in contact with each other, butwere not overconfluent. Each field was scanned to generate a digital image of fluorescence (A and B, prebleach). After the initial scan, selected cellswere photobleached (A and B, 0 minutes, numbers 1 to 9). Subsequently, three sequential scans were done at 30-second intervals to detect recoveryof fluorescence in the bleached cells (A and B, 1 minute, numbers 1 to 9). Images were digitally recorded for analysis. Several unbleached cells werealso monitored to provide control data (A and B, number 10). Typical plots of fluorescence recovery after photobleaching are shown (A and B,% prebleach vs. time). A positive slope indicates the recovery of fluorescence. The percent recovery of fluorescence over time was determinedfor each selected cell and the data were corrected for the background loss of fluorescence in one area (number 10). Original magnification �400.


Cx43 immunoreactive protein (molecular wt 41,000 to43,000) were observed in all samples, with a faster migrat-ing band (P0) and two slower migrating bands that repre-sented the two phosphorylated forms (Cx43P1 andCx43P2; Fig. 4A). Incubation with 50 or 90 mmol/L glu-cose resulted in a significant decrease of total Cx43 (Fig.4B). This was associated with a decrease of both P1 andP2 (Fig. 4C), while there appeared to be no significantchange of the P0 level (data not shown). After incubationwith 6 mmol/L HMBA for three days, total Cx43 proteinwas increased, as were both P1 and P2 when comparedwith cells that had been incubated in the absence ofHMBA (Fig. 4 B, C).

Effects of high glucose and HMBA on Cx43 mRNA

The effect of glucose on Cx43 mRNA expression wasinvestigated by semiquantitative RT-PCR analysis. Toavoid an amplification plateau, the range of cycles overwhich there is linear accumulation of PCR products mustbe determined. We therefore carried out preliminaryexperiments to determine the number of cycles thatachieved submaximal amplification of each of the genesstudied, and found that it was 23 cycles for both theCx43 gene and the housekeeping gene (GAPDH). Atypical gel is shown in Figure 5A. There was a significantdecrease of Cx43 mRNA intensity after six days of incu-bation at the higher glucose concentrations (50 or 90mmol/L) in the absence of HMBA, while the additionof 6 mmol/L HMBA to cultures with 50 or 90 mmol/Lglucose led to significant up-regulation of Cx43 mRNAafter three days (Fig. 5B).

Effects of protein kinase inhibitors on highglucose-induced changes of GJIC, as well asexpression and phosphorylation of Cx43

Two protein kinase inhibitors (calphostin C andPD98059) and an activator (TPA) were used to deter-

taining glucose (10, 50, or 90 mmol/L) or with 10 mmol/L glucose plusmannitol (40 or 80 mmol/L) to provide an equivalent osmotic load(protocol A). Symbols are: 10 mmol/L glucose (�), 50 mmol/L glucose(�), 90 mmol/L glucose (�), 40 mmol/L mannitol (�), and 80 mmol/Lmannitol (�). (B) Effect of HMBA on gap junctional intercellularcommunication (GJIC) in HPMCs exposed to high glucose. HPMCswere incubated in medium containing 10, 50, or 90 mmol/L glucose forsix days, and HMBA (1 or 6 mmol/L) was added to some cultures forthe final three days (protocol B). (C) Influence of high glucose on thedye transfer rate in HPMCs pretreated with HMBA. HPMCs wereincubated in medium containing 0 or 6 mmol/L HMBA for six days,and glucose (50 or 90 mmol/L) was added to some cultures for the finalthree days (protocol C). (A–C) The percentage of recovery rate wascalculated relative to that for cells incubated in medium containing 10mmol/L glucose without HMBA (100%). Values are the mean � SEM

Fig. 2. Effect of glucose and hexamethylene bisacetamide (HMBA) of five separate experiments performed with HPMCs isolated from fiveon the dye transfer rate in HPMCs assessed by the FRAP assay. (A) different donors (20 to 25 cells were evaluated in each experiment).Time- and dose-dependence of the dye transfer rate in HPMC exposed NS, not significant; *P � 0.05; **P � 0.01 versus cells incubated withto glucose and mannitol. HPMCs were incubated with medium con- 10 mmol/L glucose without HMBA, unless otherwise indicated.


Fig. 3. Immunostaining of human peritoneal mesothelial cells (HPMCs) for connexin 43 (Cx43). HPMCs were incubated for six days in mediumcontaining 10, 50, or 90 mmol/L glucose with or without HMBA (1 or 6 mmol/L) for the final three days (protocol B). Cx43 was detected as greenspots by indirect immunofluorescence and cell nuclei were counterstained with propidium iodide. (a, d, and g) incubation with 10 mmol/L glucose.(b, e, and h) With 50 mmol/L glucose. (c, f, and i) With 90 mmol/L glucose. (a–c) Without HMBA. (d–f) With 1 mmol/L HMBA. (g–i) With 6mmol/L HMBA. Scale bar � 50 �m.

mine whether the modification of protein kinase activity thelial cells with high concentrations of glucose. Thereduction of GJIC that we observed was accompaniedhad any effect on GJIC and/or Cx43 phosphorylationby transcriptional, translational, and post-translationalin our various assay systems. Neither calphostin C nordown-regulation of Cx43 activity. The extent of GJICPD98059 appeared to restore GJIC activity in glucose-in cells incubated with each concentration of glucoseinhibited HPMC, and the phosphorylation (P1 P2) ofappeared to be closely related to the amount and distri-Cx43 remained unchanged whether or not these inhibi-bution of Cx43 protein that could be visualized by immu-tors were present when the cells were incubated undernofluorescence or Western blotting. Pelin et al have re-high glucose conditions. TPA treatment for one hourported that Cx43 is the dominant gap junction proteinled to an almost complete disappearance of the ordinaryin primary human mesothelial cells, as we also found inCx43 bands (P0 and P1), while more slowly migratingour previous study [28, 47]. Taken together, these resultsbands corresponding to hyperphosphorylated Cx43 pro-suggest that Cx43 is the primary target when GJIC intein became apparent (Fig. 6 B, C). These changes wereHPMCs is inhibited by glucose.accompanied by a prominent reduction of the dye trans-

Other factors may be involved in the inhibition offer rate (Fig. 6A).GJIC by high glucose. However, it seems unlikely thatdirect toxicity of glucose is responsible for the decline

DISCUSSION of GJIC because no significant reduction of viability wasTo our knowledge, this is the first study to show that detected after HPMCs were incubated with high glucose

for six days, although the dye transfer rate decreased inGJIC is decreased by incubating human peritoneal meso-


Fig. 5. Effect of HMBA on Cx43 mRNA expression at various glucoselevels. (A) Cx43 mRNA was identified by RT-PCR. An equal volumeof mRNA (5 �L) was loaded onto each lane. This is a typical gelshowing the Cx43 band (506 bp) and the GAPDH band (933 bp) usedas the internal control. (B) Analysis of the change in Cx43 mRNAexpression caused by incubation with 6 mmol/L HMBA at variousglucose levels. Symbols are: (�) 10 mmol/L glucose; ( ) 50 mmol/Lglucose; ( ) 90 mmol/L glucose. The control value was arbitrarily setat one. Values are the mean � SEM of five separate experimentsperformed with HPMCs isolated from five different donors. *P � 0.05;**P � 0.01 versus incubation in 10 mmol/L glucose without HMBA,unless otherwise indicated.

a time-dependent fashion during incubation. Inoguchi etal reported that high glucose induces elevation of thediacylglycerol level in vascular tissues, with subsequentdown-regulation of GJIC occurring after incubation forthree days [48]. The time-dependent and delayed reduc-tion of GJIC in HPMCs suggests that a series of meta-

Fig. 4. Effect of HMBA on Cx43 protein expression at various glucose bolic reactions activated by high glucose may be in-levels. (A) Identification of Cx43 protein by Western blotting. Whole cell

volved. On the other hand, it was considered importantextracts (20 �g) were probed with an antibody for Cx43 as described inthe Methods section. This is a typical membrane showing three Cx43 to exclude the possibility that the effects we observedprotein bands and the �-tubulin band (internal control). (B) The relative were simply due to an increase of the osmotic load, sinceintensity of total Cx43 protein. (C) Assessment of P1 P2. The control

a high glucose concentration leads to an increased trans-value was arbitrarily set at one. Symbols in panels B and C are: (�) 10mmol/L glucose; ( ) 50 mmol/L glucose; ( ) 90 mmol/L glucose. Results membrane osmotic gradient. It has been reported thatare the mean � SEM of five separate experiments performed with HPMCs

other non-metabolized osmotic solutes, especially man-isolated from five different donors. *P � 0.05; **P � 0.01 versus incuba-tion in 10 mmol/L glucose without HMBA, unless otherwise indicated. nitol, cause dose-dependent suppression of the growth


of mesothelial cells [27]. However, we found that theeffect of equimolar mannitol on dye transfer was weakerthan that of glucose, thereby indicating that the osmoticload associated with high glucose is not likely to be amajor reason for the decline of GJIC.

To our knowledge, this is the first study to show thatincubation with high glucose causes a decrease of GJICthrough hypophosphorylation of Cx43. Our results aredifferent from those of previous studies performed onother kinds of cultured cells [23, 24, 32], which showedthat GJIC was disrupted by hyperphosphorylation ofCx43 via PKC under high glucose conditions. As ex-pected, incubation with TPA, which is another potentialactivator of PKC, caused an immediate decrease of dyetransfer by GJIC that was accompanied by hyperphos-phorylation of Cx43. However, the same kinds of hyper-phosphorylated bands as those induced by the TPA treat-ment were not seen after incubation with high glucosefor three days. Furthermore, there was no significantimprovement of dye transfer by GJIC after the additionof calphostin C, a PKC inhibitor. Matesic et al reportedtwo possible modes of Cx43 phosphorylation after theaddition of two different promoters to cultured cells, one(in which GJIC is inhibited through promotion of thePKC) induced hyperphosphorylation of Cx43 and another(that does not involve PKC) induced hypophosphoryla-tion of Cx43 [49]. These findings suggest that PKC activa-tion might not be the dominant mechanism modulatingGJIC in mesothelial cells.

In our study, high glucose reduced the phosphoryla-tion of Cx43 protein after a longer incubation period ofup to six days, and this change was closely related withthe reduction of dye transfer by GJIC. In contrast, short-term incubation with glucose for less than one hour didnot cause any detectable change of dye transfer (datanot shown). Nishikawa et al [50] and Gainer and Murray[51] reported that PKC activity could be down-regulatedthrough prolonged exposure to TPA. Hayashi et alshowed that the pattern of Cx43 phosphorylation couldchange over time in a rat liver epithelial cell line [44].These findings suggest that our results derived from long-term incubation with high glucose for several days mayhave to be distinguished from the short-term effects seenafter a number of minutes.

Fig. 6. Effects of protein kinase inhibitors on high glucose-inducedchanges of GJIC, as well as expression and phosphorylation of Cx43.(A) Effect of calphostin C, PD98059 and TPA on GJIC in HPMCsexposed to high glucose. HPMCs were incubated in medium containing10 (�) or 90 ( ) mmol/L glucose for three days and one of the inhibitors

in the Methods section. This is a typical membrane showing Cx43or the activator (calphostin C: 50 nmol/L, PD98059: 20 �mol/L andTPA: 15 ng/mL) was added to some cultures for the final four hours protein bands and the �-tubulin band (internal control). (C) Assessment

of P1 P2. The control value was arbitrarily set at one. Results are(one hour only for TPA, protocol D). The percent recovery rate wascalculated relative to that for cells incubated in medium containing 10 the mean � SEM of five separate experiments performed with HPMCs

isolated from five different donors; symbols are the same as in panelmmol/L glucose without drugs (100%). Values are the mean � SEM offive separate experiments performed with HPMCs isolated from five A. (A and C) NS, not significant; **P � 0.01 versus cells incubated

with 10 mmol/L glucose without drugs, unless otherwise indicated. Ab-different donors (20 to 25 cells were evaluated in each experiment).(B) Identification of the Cx43 protein by Western blotting. Whole cell breviations are: (�) 90 mmol/L glucose without drugs; CC, calphostin

C; PD, PD98059.extracts (20 �g) were probed with an antibody for Cx43 as described


The MEK/ERK cascade is known as the classical thelial cell repair mechanisms that are damaged by peri-MAPK cascade [29]. Cx43 is a MAPK substrate and phos- toneal dialysis. This study also showed that primary cul-phorylation of Cx43 can lead to down-regulation of GJIC tured HPMCs may be useful for investigating the effects[23, 24, 31–33]. In the present study, however, we could of high glucose dialysis fluids on GJIC in mesothelialnot find hyperphosphorylation of Cx43 under high glu- cells. An increased understanding of the mechanisms caus-cose conditions. The effect of high glucose on phosphory- ing mesothelial cell damage may contribute to the devel-lation of Cx43 could not be mimicked by the addition opment of new peritoneal dialysis fluids with significantlyof TPA, a MAPK activator. Furthermore, there was no better biocompatibility. Several safer analogues of HMBAsignificant improvement of dye transfer by GJIC after (second-generation hybrid polar compounds) have al-addition of PD98059, a MEK-inhibitor. We could find ready been investigated [37]. If any of these analoguesno evidence for a role of the MEK/ERK cascade in also can up-regulate GJIC, such drugs may be of somedown-regulating GJIC in HPMCs, at least during long- utility for the maintenance of communication betweenterm incubation with high-glucose medium. peritoneal mesothelial cells in vivo. Therefore, we are

The present study also showed that HMBA could up- planning to investigate the effectiveness of such com-regulate GJIC in mesothelial cells exposed to both nor- pounds in the future.mal and high glucose conditions. Moreover, the increaseof GJIC with HMBA pretreatment was reversed by addi- ACKNOWLEDGMENTStion of 50 or 90 mmol/L glucose. Thus, the mechanism

This study was supported in part by a gift from Mrs. Loretta Vanby which high glucose and HMBA act on GJIC in meso- Camp. Part of this work was presented as a poster at the 32nd andthelial cells might be the same. It has been shown that 33rd Annual Meetings of the American Society of Nephrology (Miami,

FL, USA, J Am Soc Nephrol 10:331A, 1999; and Toronto, Ontario,the site and extent of Cx43 phosphorylation can affectCanada, J Am Soc Nephrol 11:312A, 2000). The authors also thankthe trafficking, assembly, and activity of this protein [16]. the surgical team at Hiroshima Prefectural Hospital (Drs. Issei Tanaka,

It may well be that dephosphorylation of Cx43 by high Yasuhiko Fukuda, Tsuneo Tanaka, Shuji Hoshino, Naoki Kagawa,and Naofumi Shigeta) for providing omental samples, as well as Dr.glucose causes a reduction of GJIC along with a decreaseKiyohiko Dohi. Finally, the authors wish to thank Dr. Nicholas Topleyof Cx43 protein at gap junctions and that phosphoryla-(University of Wales College of Medicine), Dr. Donald G. MacPhee

tion of Cx43 by HMBA leads to enhancement of GJIC (Radiation Effects Research Foundation), and Dr. Hideki Kawanishi(Tsuchiya General Hospital) for their valuable contributions duringalong with an increase of this protein at gap junctions.the preparation of this manuscript.Hexamethylene bisacetamide is one of the many dif-

ferentiation-inducing agents used to control tumors by Reprint requests to Takahiko Ogawa, M.D., Ph.D., Nephrology anddifferentiation therapy [36]. It has previously been shown Dialysis Division, Hiroshima Prefectural Hospital, 1-5-54, Ujina-Kanda,

Minami-ward, Hiroshima, 734-8530, Japan.that several differentiation-inducing agents, such asE-mail: [email protected] reductase inhibitors [22] and some (but not

all) retinoids [52], can up-regulate gap junction activity.In a previous study, we also found that HMBA could APPENDIXenhance GJIC in HPMCs [28]. We were unable to detect

Abbreviations used in this article are: Cx43, connexin 43; ERK,any obvious effect of HMBA on cell cycle progression extracellular signal-regulated kinase; FCS, fetal calf serum; GJIC, gap

junctional intercellular communication; HMBA, hexamethylene bisaceta-in confluent mesothelial cells by flow cytometry (datamide; HMG-CoA, 3-hydroxy-3-methylglutaryl coenzyme A; HPMCs,not shown), although it has been reported that HMBAhuman peritoneal mesothelial cells, MAPK, mitogen-activated proteinis able to suppress the proliferation of HPMCs [38] and kinase; MEK, MAPK/ERK kinase; PBS, phosphate buffered saline;

certain other kinds of cells [36, 37]. Given these observa- PKC, protein kinase C; RT-PCR, reverse transcription-polymerase chainreaction; TPA, 12-O tetradecanoylphorbol-13-acetate; WST-1, 4-[3-(iodo-tions, it appears unlikely that the up-regulation of GJICphenyl)-2-(4-nitrophenyl)-2H-5-tetrazolio]1,3-benzene disulfonate.in mesothelial cells by HMBA is solely a function of

cell cycle-dependent processes. Although HMBA showsREFERENCESlittle toxicity for various types of cultured cells at concen-

1. Heimburger O, Waniewski J, Werynski A, et al: Peritoneal trans-trations of 10 mmol/L or less [53] and levels of at leastport in CAPD patients with permanent loss of ultrafiltration capac-6 mmol/L appear to be safe for HPMCs [38], we have noity. Kidney Int 38:495–506, 1990

information as to whether or not prolonged exposure to 2. Schneble F, Bonzel KE, Waldherr R, et al: Peritoneal morphol-HMBA can cause mesothelial cell damage in vivo. ogy in children treated by continuous ambulatory peritoneal dial-

ysis. Pediatr Nephrol 6:542–546, 1992A reduction of normal cell-to-cell communication may3. Nagy JA: Peritoneal membrane morphology and function. Kidneyinterfere with the usual mechanism of mesothelial cell Int 50(Suppl 56):S2–S11, 1996

repair [13]. This study confirmed that HMBA can block 4. Simionescu M, Simionescu N: Organization of cell junctions in theperitoneal mesothelium. J Cell Biol 74:98–110, 1977the adverse effects of high glucose on GJIC in vitro, a

5. Michailova KN: A combined electron microscopic investigationfinding that may have some clinical importance in theof the peritoneal mesothelium in the rat. Eur J Morphol 33:265–277,

future. Instillation of HMBA into the peritoneal cavity 19956. Di Paolo N, Sacchi G, De Mia M, et al: Morphology of theas an up-regulator of GJIC could help to restore meso-


peritoneal membrane during continuous ambulatory peritoneal 30. Haneda M, Araki S, Togawa M, et al: Mitogen-activated proteindialysis. Nephron 44:204–211, 1986 kinase cascade is activated in glomeruli of diabetic rats and glomer-

7. Breborowicz A: In vitro study on the biocompatibility of the ular mesangial cells cultured under high glucose conditions. Diabe-peritoneal dialysis solution. Perit Dial Int 13(Suppl 2):S105–S108, tes 46:847–853, 19971992 31. Warn-Cramer B, Lampe P, Kurata W, et al: Characterization of

8. Topley N, Coles GA, Williams JD: Biocompatibility studies on the mitogen-activated protein kinase phosphorylation sites on theperitoneal cells. Perit Dial Int 14(Suppl 3):S21–S28, 1994 connexin-43 gap junction protein. J Biol Chem 271:3779–3786, 1996

9. Gotloib L, Shostak A, Wajsbrot V, et al: The cytochemical profile 32. Stalmans P, Himpens B: Effect of increasing glucose concentra-of visceral mesothelium under the influence of lactated-hyperosmo- tions and protein phosphorylation on intercellular communicationlar peritoneal dialysis solutions. Nephron 69:466–471, 1995 in cultured rat retinal pigment epithelial cells. Invest Ophthalmol

10. Gotloib L, Wajsbrot V, Shostak A, et al: Morphology of the Vis Sci 38:1598–1609, 1997peritoneum: Effect of peritoneal dialysis. Perit Dial Int 15(Suppl 33. Warn-Cramer B, Cottrell G, Burt J, et al: Regulation of con-7):S9–S12, 1995 nexin-43 gap junctional intercellular communication by mitogen-

11. Topley N, Mackenzie R, Petersen MM, et al: In vitro testing activated protein kinase. J Biol Chem 273:9188–9196, 1998of a potentially biocompatible continuous ambulatory peritoneal 34. Stylianou E, Jenner LA, Davies M, et al: Isolation, culture anddialysis fluid. Nephrol Dial Transplant 6:574–581, 1991 characterization of human peritoneal mesothelial cells. Kidney Int

12. Topley N, Petersen MM, Mackenzie R, et al: Human peritoneal 37:1563–1570, 1990mesothelial cell prostaglandin synthesis: Induction of cyclooxygen- 35. Lever JE: Chemical inducers of differentiation in a long-term renalase mRNA by peritoneal macrophage-derived cytokines. Kidney cell line. Environ Health Perspect 80:173–180, 1989Int 46:900–909, 1994 36. Marks PA, Richon VM, Kiyokawa H, et al: Inducing differentia-

13. Yung S, Davies M: Response of the human peritoneal mesothelial tion of transformed cells with hybrid polar compounds: A cellcell to injury: An in vitro model of peritoneal wound healing. cycle-dependent process. Proc Natl Acad Sci USA 91:10251–10254,Kidney Int 54:2160–2169, 1998 1994

14. Charles AC, Naus CCG, Zhu D, et al: Intercellular calcium signal- 37. Richon VM, Webb Y, Merger R, et al: Second generation hybriding via gap junctions in glioma cells. J Cell Biol 118:195–201, 1992 polar compounds are potent inducers of transformed cell differenti-

15. Sammak PJ, Hinman LE, Tran POT, et al: How do injured cells ation. Proc Natl Acad Sci USA 93:5705–5708, 1996communicate with the surviving cell monolayer? Metabolic traf- 38. Schroppel B, Fischereder M, Wiese P, et al: Expression of glucoseficking through astrocytic gap junctions. J Cell Sci 110:465–475, transporters in human peritoneal mesothelial cells. Kidney Int 53:1997 1278–1287, 1998

16. Trosko JE, Ruch RJ: Cell-cell communication in carcinogenesis. 39. Kobayashi E, Nakano H, Morimoto M, et al: Calphostin C (UCN-Front Biosci 3:208–236, 1998 1028C), a novel microbial compound, is a highly potent and specific

17. Loewenstein WR: Junctional intercellular communication and the inhibitor of protein kinase C. Biochem Biophys Res Commun 159:control of growth. Biochim Biophys Acta 560:1–65, 1979 548–553, 1989

18. Bruzzone R, White TW, Paul DL: Connections with connexins: 40. Alessi D, Cuenda A, Cohen P, et al: PD 098059 is a specificThe molecular basis of direct intercellular signaling. Eur J Biochem inhibitor of the activation of mitogen-activated protein kinase ki-238:1–27, 1996nase in vitro and in vivo. J Biol Chem 270:27489–27494, 199519. Trosko JE, Yotti LP, Warren ST, et al: Inhibition of cell-cell

41. Sitter T, Mandl-Weber S, Wornle M, et al: D-glucose increasescommunication by tumor promoters. Carcinogenesis 7:565–585, 1982the synthesis of tissue-type plasminogen activator (t-PA) in human20. Trosko JE, Jone C, Chang CC: Inhibition of gap junctional-medi-peritoneal mesothelial cells. Thromb Haemost 82:1171–1176, 1999ated intercellular communication in vitro by Aldrin, dieldrin and

42. Ishiyama M, Tominaga H, Shiga M, et al: A combined assay oftoxaphene: A possible cellular mechanism for their tumor-promot-cell viability and in vitro cytotoxicity with a highly water-solubleing and neurotoxic effects. Mol Toxicol 1:83–93, 1987tetrazolium salt, neutral red and crystal violet. Biol Pharm Bull21. Rosenkranz HS, Pollack N, Cunningham AR: Exploring the19:1518–1520, 1996relationship between the inhibition of gap junctional intercellular

43. Wade MH, Trosko JE, Schindler M: A fluorescence photobleach-communication and other biological phenomena. Carcinogenesising assay of gap junction mediated communication between human21:1007–1011, 2000cells. Science 232:525–528, 198622. Ruch RJ, Madhukar BV, Trosko JE, et al: Reversal of ras-induced

44. Hayashi T, Matesic DF, Nomata K, et al: Stimulation of cellinhibition of gap-junctional intercellular communication, transfor-proliferation and inhibition of gap junctional intercellular commu-mation, and tumorigenesis by lovastatin. Mol Carcinog 7:50–59,nication by linoleic acid. Cancer Lett 112:103–111, 19971993

45. Crow DS, Beyer EC, Paul DL, et al: Phosphorylation of con-23. Inoguchi T, Ueda F, Umeda F, et al: Inhibition of intercellularnexin43 gap junction protein in uninfected and Rous sarcoma virus-communication via gap junction in cultured aortic endothelial cellstransformed mammalian fibroblasts. Mol Cell Biol 10:1754–1763,by elevated glucose and phorbol ester. Biochem Biophys Res Com-1990mun 208:492–497, 1995

46. Trosko JE, Chang CC, Wilson MR, et al: Gap junctions and24. Kuroki T, Inoguchi T, Umeda F, et al: High glucose induces alter-regulation of cellular functions of stem cells during developmentation of gap junction permeability and phosphorylation of con-and differentiation. Methods 20:245–264, 2000nexin-43 in cultured aortic smooth muscle cells. Diabetes 47:931–

47. Pelin K, Hirvonen A, Linnainmaa K: Expression of cell adhesion936, 1998molecules and connexins in gap junctional intercellular communi-25. Lorenzi M, Cagliero E, Toledo S: Glucose toxicity for humancation deficient human mesothelioma tumor cell lines and commu-endothelial cells in culture: Delayed replication, disturbed cell cy-nication competent primary mesothelial cells. Carcinogenesis 15:cle, and accelerated death. Diabetes 34:621–627, 19852673–2675, 199426. Lorenzi M, Montisano DF, Toledo S, et al: High glucose induces

48. Inoguchi T, Xia P, Kunisaki M, et al: Insulin’s effect on proteinDNA damage in cultured human endothelial cells. J Clin Investkinase C and diacylglycerol induced by diabetes and glucose in77:322–325, 1986vascular tissues. Am J Physiol 267(3 Pt 1):369–379, 199427. Breborowicz A, Rodela H, Oreopoulos DG: Toxicity of osmotic

49. Matesic DF, Rupp HL, Bonney WJ, et al: Changes in gap-junctionsolutes on human mesothelial cells in vitro. Kidney Int 41:1280–permeability, phosphorylation, and number mediated by phorbol1285, 1992ester and non-phorbol-ester tumor promoters in rat liver epithelial28. Ogawa T, Hayashi T, Kyoizumi S, et al: Up-regulation of gapcells. Mol Carcinog 10:226–236, 1994junctional intercellular communication by hexamethylene bisacet-

50. Nishikawa K, Yamamoto S, Nagumo H, et al: The presence ofamide in cultured human peritoneal mesothelial cells. Lab Investphorbol ester responsive and non-responsive forms of the zeta iso-79:1511–1520, 1999zyme of protein kinase C in mouse epidermal cells. Cell Signal 7:29. Seger R, Krebs EG: The MAPK signaling cascade. FASEB J

9:726–735, 1995 491–504, 1995


51. Gainer HSC, Murray AW: Diacylglycerol inhibits gap junctional hanced growth control and cancer prevention, in Gap Junctions,edited by Hall JE, Zampighi GA, Davis RM, Amsterdam, Else-communication in cultured epidermal cells: Evidence for a role of

protein kinase C. Biochem Biophys Res Commun 126:1109–1113, vier Science Publishers Ltd., 1993, pp 301–30953. Yet SF, Kong CT, Peng H, et al: Regulation of Na/glucose co-1985

52. Hossain MZ, Zhang LX, Bertram JS: Retinoids and carotenoids transporter (SGLUT 1) mRNA in LLC-PK1 cells. J Cell Physiol158:506–512, 1994upregulate gap-junctional communication: Correlation with en-

hexamethylene bisacetamide protects peritoneal mesothelial cells from glucose

Documents