Journal of Cellular Biochemistry 99:435–449 (2006) Reduction of Anabolic Signals and Alteration of Osteoblast Nuclear Morphology in Microgravity Millie Hughes-Fulford,1,2,3* KarstenRodenacker,4and Uta Ju¨tting4 1Northern California Institutefor Research andEducation, San Francisco, California 2DepartmentofMedicine, University ofCalifornia, San Francisco,California 3DepartmentofVeteran Affairs, LaboratoryofCell Growth and Differentiation,San Francisco, California 4GSF—National Research Centerfor Environment and Health,Institute ofBiomathematicsand Biometry, Ingolsta¨dter Landstraße1, D-85764 Neuherberg, Germany Abstract Bone loss has been repeatedly documented in astronauts after flight, yet little is known about the mechanism of bone loss in space flight. Osteoblasts were activated during space flight in microgravity (mg) with and withouta1gravity(1g)fieldand24geneswereanalyzedforearlyinduction.Inductionofproliferatingcellnuclearantigen (PCNA), transforming growth factor b (TGFb), cyclo-oxygenase-2 (cox-2), cpla2, osteocalcin (OC), c-myc, fibroblast growthfactor-2(fgf-2),bcl2,bax,andfgf-2messageaswellasFGF-2proteinweresignificantlydepressedinmgwhen comparedtoground(gr).Artificialonboardgravitynormalizedtheinductionofc-myc,cox-2,TGFb,bax,bcl2,andfgf-2 messageaswellasFGF-2proteinsynthesisinspaceflightsamples.Innormalgravity,FGF-2inducesbcl2expression;we foundthatbcl2expressionwassignificantlyreducedinmicrogravityconditions.Sincenuclearshapeisknowntoelongate intheabsenceofmitogenslikeFGF-2,weusedhigh-resolutionimage-basedmorphometrytocharacterizechangesin osteoblast nuclear architecture under microgravity, 1 g flight, and ground conditions. Besides changes in cell shape (roundish/elliptic),otherhigh-resolutionanalysesshowclearinfluencesofgravityontheinnernuclearstructure.These changesoccurinthetexture,arrangement,andcontrastofnuclearparticlesandmathematicalmodelingdefinesthesingle cell classification of the osteoblasts. Changes in nuclear structure were evident as early as 24 h after exposure to microgravity. This documented alteration in nuclear architecture may be a direct result of decreased expression of autocrine and cell cycle genes, suggesting an inhibition of anabolic response in mg. Life on this planet has evolved inanormalgravityfieldandthesedatasuggestthatgravityplaysasignificantroleinregulationofosteoblasttranscription. J.Cell.Biochem.99:435–449,2006. (cid:1)2006Wiley-Liss,Inc. Keywords: microgravity;growthregulation;cytometry;morphometry;photometry;fgf-2;bcl2 Aprojectedbonelossof20–30%inastronauts and human studies, investigators found that isoneofthemajorphysiologicalshowstoppers loss of weight bearing bone can be as high as in the proposed 30-month manned mission to 3% per month andthis loss isprimarilydueto Mars.Forthefirstthreedecadesofspaceflight, lack of new osteoblast growth in space flight scientists have discovered systemic changes in [Wronski and Morey, 1983; Nicogossian et al., the body, which included loss of bone and in- 1989; Marie et al., 2000; Vico et al., 2000]. creasesinurinarycalcium.Fromtheseanimal Although flight opportunities have been lim- ited, the volume of data about microgravity effectsoncellproliferationanddifferentiationis accumulating. It has been demonstrated that cellgrowthanddifferentiationofosteoblastsis Grant sponsor: NASA; Grant numbers: NAG-2-1086, inhibited or altered in a microgravity environ- NAG-2-1286, NCC-2-136; Grant sponsor: Department of ment [Fitzgerald and Hughes-Fulford, 1996; VeteransAffairsResearchServiceMeritReview. Hughes-Fulford and Lewis, 1996; Carmeliet *Correspondenceto:MillieHughes-Fulford,PhD,Director, et al., 1998; Landis et al., 2000; Granet et al., Lab of Cell Growth, VA Medical Center Mail Code 151F, 2001] suggesting that the osteoporosis seen in 4150ClementStreet,SanFrancisco,CA94121. E-mail:[email protected] spaceflight is occurring at the cellular level. Received12January2006;Accepted15February2006 Alteredgeneexpression,reductionintranscrip- DOI10.1002/jcb.20883 tion factors, and reductions in growth factor (cid:1)2006Wiley-Liss,Inc. 436 Hughes-Fulford et al. related proteins have also been observed in portion of the cells exhibited a spindle shape othercellsinmicrogravity[deGrootetal.,1990; after 4 days in microgravity, with the most CubanoandLewis,2000;Goodwinetal.,2000]. dramatic change observed being the elongation In2001,Lewisetal.founda5.9-foldreductionin of nuclear shape [Fitzgerald and Hughes-Ful- message for TGFa in space-flown lymphocytes ford,1996]. [Lewisetal.,2001]. Since it has been suggested that changes in Inthisreport,wefoundsignificantreductionin nucleararchitectureandcontrolofgeneexpres- expression ofseveral osteoblast genesincluding sionareinterrelated[Debnathetal.,1999],we fibroblast growth factor-2 (fgf-2), cyclo-oxyge- analyzed message expression of other genes nase-2(cox-2),c-myc,transforminggrowthfactor involvedincellgrowthandcorrelatedmorpho- b (TGFb), proliferating cell nuclear antigen logical changes between osteoblasts grown in (PCNA),osteocalcin(OC),andbaxinosteoblasts microgravity,1gorgr.Inthisstudywereport grown in mg. Since fgf-2, TGFb, and cox-2 are significant microgravity-induced alterations osteoblastautocrine/paracrinefactorsimportant in transcription factor message and protein intheG1-Sphasetransition[Canalisetal.,1991; expression accompanied by dramatic changes Hughes-Fulford et al., 1992; Stachowiak et al., in osteoblast cell morphology. The changes in 1994;Coffinetal.,1995;KeresztesandBoonstra, gene expression observed in osteoblasts in mg 1999],itispossiblethattheirdecreasedexpres- werenotobservedinosteoblastsmaintainedin sionmightbelinkedwithinhibitionofactivation anin-flight1gfield,implyingagravityrequire- ofbonegrowthinmicrogravity. ment for normal osteoblast growth. These Since fgf-2 is a known osteoblast autocrine/ combinedchangesinosteoblastcellphysiology paracrine factor important in the G1-S phase suggest thatthe decreasein bone formationin transition[Canalisetal.,1991;Stachowiaketal., astronautsmaybeinpartduetomicrogravity- 1994;Coffinetal.,1995;KeresztesandBoonstra, inducedalterationsintheosteoblastatthegene 1999], we hypothesized that changes in fgf-2 expressionandsignaltransductionlevel. expression may be linked to inhibition of bone growth in mg. In this study, two groups of MATERIALSAND METHODS osteoblasts onboard the Space Shuttle were Preparation of Cells exposedtomgfor19hbeforetheywerestimulated withfetalcalfsera(FCS);onegroupwasgrownin The MC3T3-E1 osteoblast cell line clonally microgravity, the other in an artificial 1 g derived from embryonic mouse calvaria environment.Weanalyzedexpressionoffgf-2,a [Hiramatsu et al., 1983], was kindly provided mitogen,andtranscriptionfactorthatpromotes to us by Dr. M. Kumegawa (Josai Dental proliferation. University, Japan). The cell line was main- As part of our study of fgf-2 induction, we tainedatlowpassagenumber.Cellsweregrown examined concomitant changes in nuclear mor- in alpha minimal essential medium (aMEM) phology of osteoblasts grown in microgravity. with 10% fetal calf serum (Hyclone Labs, Inc., Rijken[deGrootetal.,1991]firstnoticedchanges Logan, UT) supplemented with 2 mM L-gluta- in cell shape in response to changed gravity in mine (Sigma, St. Louis, MO), 25 mM HEPES, soundingrocketexperimentsin1991.Morpholo- and antibiotic–antimycotic solution (100 U gicchangeswerealsoreportedinparabolicflight penicillin/ml,0.01 mgstreptomycin/ml,0.25mg with osteoblasts exhibiting cytoplasmic retrac- amphotericinB/ml).Cellsweregrownina378C tion and membrane ruffling in ROS/17/2.8 cells incubator with 5% CO . They were fed three 2 [Guignandon et al., 1995]. Increased PGE was times a week and passaged at confluence. For 2 alsofoundinmedia,accompaniedbysignificant spaceflight experimental samples, 120,000– flight-induced changes that included a decrease 200,000 cells were plated onto non-coated, in cell area and irregular shape in some cells sterile 11(cid:1)22 mm glass coverslips (Thomas [Guignandonetal.,1995].Indomethacinreduced Scientific, Swedesboro, NJ), placed in 6-well the irregular shape changes, but not the loss of plates, and grown in 10% serum-containing cell area. Osteoblasts grown in microgravity on aMEM overnight. Cell-coated coverslips were STS-56 demonstrated a reduction of growth in then transferred into 2% serum-containing osteoblasts activated in-flight, suggesting that mediumforspaceflightinplungerboxunits.In microgravitycausedadecreaseincellprolifera- accordancewithNASAflightrules,theplunger- tion resulting in changes in cytoskeleton. A box units were held for 17 h in the Shuttle at Reduction of Anabolic Signals 437 mid-deck temperature prior to launch. This, CentrumvoorConstructieMechatronica(CCM; combined with low serum-containing medium, Neuenen, Netherlands). The hardware con- placedtheosteoblastinaquiescentstatepriorto sisted of the CCM plungerbox and its Type I Space Shuttle launch. In order to achieve our containerdevelopedforspaceflightcellculture. goaltostudychangesinnuclearmorphologyin The Type I container provided a second level microgravity, we launched cells in a quiescent of fluid containment. The plungerboxes were condition. Osteoblasts were sera activated on- designed to provide a sterile environment for orbit and mg or 1 g samples collected in-flight cell growth activation and fixation in a micro- (Fig.1).Therewerefoursamples(n¼4)foreach gravity condition. The plungerbox is composed time point in mg, 1 g flight, and gr samples. oftwoindependentculturechambersthateach NASA’sLSLE(LifeSciencesLaboratoryEquip- holdtwo11mm(cid:1)22mmglasscoverslips.Each ment) freezer, was used for sample storage on culture chamber was a separate sample. For the Space Shuttle. Supplements, such as dex- each condition a sample size of n¼4 indepen- amethasone b-glycerol phosphate and ascorbic dent cultures was used. Each separate culture acid were not added to the media since these chamber has a series of reservoirs filled by agentsareknowntodirectlyeffectgeneexpres- either10%serum-containingaMEMorfixative, sionandcellmorphologyoftheosteoblast. whichcanbetransferredintoandoutofthecell culture compartment by manually releasing a Biorack Facilityand Osteo Hardware spring-loadedplunger. CellsweregrownintheBiorack,amultiuser Experiment Time Line facility,whichconsistsofincubatorswithvari- ablegravitycentrifuges,acooler,afreezer,and Quiescent cells were stored in the Space a sealed glovebox. Two identical Biorack Shuttlemid-decklockeruntil19hafterlaunch. modules were used: one remained on earth (gr The astronauts then transferred the samples samples) and the other was integrated into intoa378Cincubatorfor1hbeforestimulating SpaceHab and flown on each of the Space thecellstogrowinthe1gcentrifugeormgstatic Shuttle flights STS-76, STS-81, and STS-84. rackbychangingtheirmediafrom2%serumto Biorackhasanimportantadvantageoverother 10% serum-containing medium (T¼0). mg and microgravityfacilitiesinthatitprovidesasmall 1gsampleswerefixedduringflightat24hwith radius(78mm),slowrotating(107.0(cid:2)0.5rpm) 3.7%formaldehydesolutionandwerestoredat centrifuge for artificial 1 g in orbit. Because of þ58C for the remainder of the Space Shuttle their proximity, both the mg and 1 g samples mission, while the GITC samples were frozen experience identical launch vibrations, accel- untilreturn.Therewerefourseparateindepen- erations,cosmicradiation,andotherunknown dentsamplesforeachgravityconditionateach conditions of flight. In addition, an identical timepoint(Fig.1). experiment was performed in the Biorack module on earth gr samples with a 2 h delay BIOCHEMICAL MEASUREMENTS fromin-flightproceduresinordertocontrolfor FGF-2 Protein Analysis any off-nominal operations during the flight. The Osteo (the name ofthisflight experiment) Analysis of FGF-2 was completed with a spaceflighthardwarewasdesignedaccordingto QuantitiesELISAImmunoassay(R&DSystems, EuropeanSpaceAgency(ESA)specificationsfor Inc., Minneapolis, MN) according to manufac- use in the Biorack facility and constructed by turer’s instructions. Cellular protein was col- lected from the GITC subnatant as previously described[FitzgeraldandHughes-Fulford,1999]. Analysis ofGene Expression TableIshowstheoverviewofgeneexpression measured in these experiments. RNA was ex- tracted and purified by a modified acid guan- idiumthiocyanate/phenol/chloroformextraction method previously described [Fitzgerald and Fig. 1. Time line of experiment from handover. 19 h pre- Hughes-Fulford, 1996; Barry et al., 1999; launch, activation in-flight, and termination of experiment in-flight. FitzgeraldandHughes-Fulford,1999]. TABLE I. OsteoblastGeneExpressioninMicrogravity GENE STS-76,81,and84composite Timepoint #Ofcycles Present Significantchange 18s T¼0 18 Yes No T¼3 18 Yes T¼24 18 Yes CPHI T¼0 22 Yes No T¼3 24 Yes T¼24 24 Yes IL-1a T¼0 42 No — T¼3 42 No T¼24 42 No cox-2 T¼0 34 Yes Yes T¼3 30 Yes T¼24 30 Yes cox-1 T¼0 40 No — T¼3 40 No T¼24 40 No cpla2 T¼0 32 Yes Yes T¼3 32 Yes T¼24 32 No EP1 T¼0 37 Yes No T¼3 37 Yes T¼24 37 Yes EP2 T¼0 40 No — T¼3 40 No T¼24 40 No EP3 T¼0 — — T¼3 47 No T¼24 47 No LDLr T¼0 38 — — T¼3 38 No T¼24 38 No CyclinA T¼0 38 Yes No T¼3 38 Yes T¼24 38 Yes CyclinE T¼0 38 — No T¼3 38 Yes T¼24 38 Yes PCNA T¼0 42 Yes No T¼3 42 Yes T¼24 42 — c-myc T¼0 38 Yes Yes T¼3 38 Yes T¼24 38 Yes cjun T¼0 35 No — T¼3 35 No T¼24 35 No TGFb1 T¼0 38 Faint Yes T¼3 38 Yes T¼24 38 No FGF-2 T¼0 34 — Yes T¼3 34 No T¼24 34 Yes Actin T¼0 30 Yes No T¼3 30 Yes T¼24 30 Yes Fibronectin T¼0 34 Yes No T¼3 34 Yes T¼24 34 Yes Collagen T¼0 33 No — T¼3 33 No T¼24 33 No Osteocalcin T¼0 — Yes T¼3 38 Yes T¼24 38 Yes b1integrin T¼0 47 No — T¼3 47 No T¼24 47 No Bax T¼0 30 Yes Yes T¼3 30 Yes T¼24 30 Yes bcl2 T¼0 34 Yes Yes T¼3 38 Yes T¼24 38 Yes OsteoblastswereflownontheShuttleinaquiescentstate;afterinsertionintoorbit,thecultureswereactivatedwithFCS.Manyofthe genesthathadsignificantlychangedexpressioninmicrogravityareearlypleiotopicgenesthatnormallyareincreasedbygrowthfactors foundinFCS.Significantchangesatthecollectiontimepoints(Fig.1)areseeninFigures2and3(n¼4). Reduction of Anabolic Signals 439 RT-PCR Analysis Cell Morphology Osteoblasts were fixed on coverslips with The semi-quantitative RT-PCR analysis and 3.7% formaldehyde in phosphate buffered sal- primer sequences for many of the genes have ineduringspaceflightandstoredat58Cfor1– been described [Fitzgerald and Hughes-Ful- 2 weeks before staining. There was no visible ford, 1996; Barry et al., 1999; Fitzgerald and difference between control samples that were Hughes-Fulford, 1999]. To correct for small immediately stained and samples stored in variations between experiments, each PCR formaldehydeat58C.Fornuclearvisualization, productwas compared toan internal standard coverslipswerestainedwithHoechstdyenum- ofcyclophilinor18Sproductsderivedfromthe ber 33258 and rinsed in distilled water. The sameRTreaction.Sincesamplesizewassmall dried osteoblast coverslips were mounted onto (200,000cells),RNAcontentwasheldconstant slidesandphotographedwithaZeissAxioscope andlinearRT-PCR wasaccomplished by vary- microscope at 40(cid:1) and 100(cid:1) magnification. ing the number of PCR cycles. Primers were Slides were processed at the same time under designed in this laboratory to span introns to identical conditions. Photographic exposure eliminate the possibility of detection of DNA. was taken at identical exposure times and PCRconditionswereestablishedsothatampli- conditions(TablesVIandVII;Figs.4–7,9). fication reaction was stopped in the linear range. Primers not previously described are IMAGING METHODS 18s F-TCA AGA ACG AAA GTC GGA GG; R- GGA CAT CTA AGG GCA TCA CA; CPHI F- Image Acquisition CGT CTC CTT TGAG CTG TTT GCA GAC R- TableIIshowsnumberofimagesatthetotal CATAATCATAAACTAACTCTGCAATCCA number of cells evaluated. Images were digi- GC;fgf-2F-AACTACAACTTCAAGCAGAAG tized from the color film slides with an Epson AGAGAR-TTAAGATCAGCTCTTAGCAGA Expression 836XL scanner into Adobe Photo- CAT; bcl2 F ACTTGTG GCCCA GATAGG- shop(AdobeSystems,Inc.,MountainView,CA) CACCCAG; R CGACTTC GCCGAGATG (Table III and Fig. 4). Images were stored as TCCAGC CAG; IL-1b F-GTC TCT GAA TCA RGB (red, green, blue) TIF format image files GAAATCCTTCTATCR-ATGTCAAATTTC (Fig.5).FromtheseRGBimages(Fig.5a)only ACT GCT TCA TCC; cpla2 F-GGA TTC TCT theREDchannelimage(Fig.5R)wasusedsince GGTGTGATGAAGG,R-CCCAATCTGCAA thenuclearstructureisbestrepresentedbythis ACATCAGCep1F-CTGGAGGGGCGTGTC channel. ATTTC,R-GCTGCGGAGGGTGGCTGTGG ep2F-CTACGCCGCCTTCTCTTACATG,R- Image Preprocessing and Segmentation GATGTCTTTCACCACGTTTGGC;ep3CAT All image analysis was performed with IDL CACCATGATGGTCACTGGC,R-GTCTTC software(InteractiveDataLanguage,Research ATGTGGCTGGGATACC;LDLrF-CAATGT System,Inc.,Boulder,CO).Forsegmentation,a CTCACCAAGCTCT,R-TCTGTCTCGAGG globalluminancevaluewasautomaticallyesti- GGT AGC T; cyclin A F AAA CAG CCT GCC matedfromeveryimageLandusedforthresh- TTCACCATTCAR-ATATTCTTCTCCCAC oldingoftheimagecontentintobackgroundand CTCAACCA;cyclinE,F-CAAGTGGCCTAT rough cell nuclei mask M (Fig. 6). Cell cluster GTCAACGAR-ACTGCTGGGTGGGGGTGT r consisting of touching or occluding cells were CA; PCNA F-CAC CAA ATC AAG AGA AAG delineatedbymanualinteraction.Bordertouch- TT, R-GAG AAT GGA GAG AAA AGA AA; c- ingcellsweredeletedresultinginmaskM.The mycF-CCAACAGGAACTATGACCTCG,R- CCACATACAGTCCTGGATG;cjunF-GGA TABLE II. NumberofImagesandCells AACGACCTTCTATGACGATGCCCTCAA, Evaluated R-GAACCCCTCCTFCTCATCTGTCACGTT CTT; collagen, F-GTC CCC CTG GCT CTG Number Number CTG GTT, R- TTT GGG TTG TTC GTC TGT Abbreviation ofimages ofcells Description TTC; b1 Integrin, TGT TCA GTG CAG CCT gr 3 62 Normalgravity TCA,R-CCTCATACTTCGGATTGACC;bax 1g(flight) 5 88 Gravityby F-CAGCTCTGAGCAGATCATGAAGACA, centrifuge mg(flight) 3 52 Staticinflight R-GCCCATCTTCTTCCAGATGGTGAGC. 440 Hughes-Fulford et al. TABLE III. Three-ClassJackknifedClassificationResultsWiththeFour MostImportantFeaturesandTheirCorrespondingF-Values (AlsoseeTableVI) % gr 1g mg Feature F-value gr 75 47 13 2 P2A 49.7 1g 58.0 22 51 15 HETERO 28.4 mg 88.5 1 5 46 GCV 22.7 Total 71.3 RHA 11.3 cuttinglinesfortheedgeoftouchingcellswere set only those features were used in the setinthecenteroftheoverlappingoroccluding, selection steps that were considered indepen- typically lens-shaped areas. Nuclei with large dent of possible global acquisition influences overlap with others were rejected from further such as relative features and which were analysis.Forfeatureextraction,theREDchan- univariatesignificant.Typically,featuresmea- nel, image R (Fig. 5R) was transformed to the suring area or total fluorescent intensity were fluorescence density image F (Fig. 5c) by the only used after normalization, for example, formula: particle area divided by nuclear area. Due to the small number of cells, only up to seven 255(cid:3)R F ¼(cid:3)log (three-class) or six (two-class) features respec- maxð255(cid:3)RÞ tivelywerestepwiseselectedonthebasisofF- Thisisanalogoustothewell-knowntransfor- statisticstoavoidoverfitting.Thevalueforthe mationfromtransmittedlightTtoextinctionE. first chosen feature is univariate whereas the Extinctionisequivalenttotheopticaldensityin following F-values are multivariate reflecting quantitativemicroscopy. the impact of results after using this feature togetherwiththealreadyselectedfeatures.The T E¼(cid:3)log classification matrices show the results of the maxT hold-one-out (jackknifed) method because no testsetwasavailable.Allstatisticalevaluations Feature Extraction were done at the 95% confidence level. For From each cell nucleus, represented by its display of the three-class discrimination, the mask M and the corresponding region in the canonical variables are used which are linear fluorescentdensityimageF,numerousquanti- combinations of the selected features. Linear tative features were calculated describing discriminant analysis results in the so called shape, photometric, and textural properties. discriminant function defined by a ((n(cid:3)1)- The latter are understood as properties of the classes(cid:1)m-feature) matrix D. The canonical arrangement of fluorescent particles, which in variablesCANarethandefinedas thiscaseisDNA,visualizedbyHoechst33528. 8 9 feature Thesefeaturesaredescribedindetailelsewhere < 1= CAN ¼D(cid:4) (cid:4)(cid:4)(cid:4) [RodenackerandBengtsson,2003](Fig.7).For :feature ; n classification, selected features are briefly describedinTables6and7.Theyareillustrated ThisistheproductofthediscriminantmatrixD by image sequences showing the changing and the selected features. This is a method to featurevaluewithappropriatecellimages. projecttheobservationsforvisualization. RESULTS Statistics and Classification Gene Expression and Protein Levels All statistical evaluations were completed inMicrogravity using SAS (SAS Institute, Inc., Cary, NC) and BMDP(StatisticalSoftware,Inc.,LosAngeles, Weanalyzedinductionof24osteoblastgenes CA)programpackages.Allcellsfromspecimens using linear semi-quantitative RT-PCR. We belongingtothesameexperimentalconditions found reduced gene expression in 10 of these were pooled and stepwise linear discriminant genes; no change in 9 of the genes and copy analysis was applied. From the whole feature number was too low to measure in the Reduction of Anabolic Signals 441 Fig.2. Analysisofgeneexpressionat3hunder1gflight,mgflight,andground.N¼4,significancebetween 1gandmglevels:P<0.001,significanceingroundandmglevels:P<0.01(barsdenotes(cid:2)SD). remaining genes examined. Osteoblasts were genous levels of cellular fgf-2 message and activated during space flight in mg with and protein in these microgravity grown samples. without a 1 g field and results compared to gr Inourstudies,wefoundthatthefgf-2message samples. Induction of PCNA, TGFb, cox-2, induction was depressed 24 h after activation cpla2, osteocalcin, c-myc, fgf-2, bcl2, and bax (Fig. 3). Moreover, when the cellular protein were significantly depressed in mg when com- levels were measured in these cells, we found pared to gr (Figs. 2 and 3). Artificial onboard that FGF-2 protein synthesis was also signifi- gravitywasabletonormalizetheinductionofc- cantly reduced in mg flight, however, the myc,cox-2,TGFb,fgf-2,bax,andbcl2aswellas presenceofartificialgravityin-flightincreased FGF-2proteinsynthesisinspaceflightsamples. theabilityofthecellstosynthesizefgf-2(Fig.3) Therewasnochangeinquiescentosteoblast 24hafterserumactivation.Finally,changesin PGE levels (gr 415(cid:2)163 vs. mg 453(cid:2)189). bcl2 message(Fig.3)inductionparallel thatof 2 Eventhoughthecox-2messagewasreducedat fgf-2 message. The bcl2 message was signifi- 24 h, the PGE levels were high. However, in- cantly decreased in microgravity compared to 2 flight at 24 h, both the 1 g and the mg samples groundand1gsamples. had a significant rise in PGE levels: 1 g 2 Three-Class Case 6,065(cid:2)870 and mg 6,959(cid:2)1132 versus gr 3,904(cid:2)1673(P<0.005). This results in the most discriminative fea- Since it has been previously demonstrated turesP2A,HETERO,GCV,RHA,RHUA,NR1, that a lack of mitogen causes elongation of and HLM3. A short description of the features nuclei [Foote et al., 1996], we measured endo- selected is given in Tables 6 and 7. Cell nuclei 442 Hughes-Fulford et al. Fig. 3. Analysisofgeneexpressionat24hunder1gflight,mgflight,andground.N¼4,significance between1gandmglevels:P<0.001,significanceingroundandmglevels:P<0.01(barsdenotes(cid:2)SD). showcleardifferencesinshapeandtexture.The ing and the less pronounced change between firstselectedfeature(P2A)isshape,remaining background inside the nucleus and bright features are all texture related. Variation of particle(GCV). P2Aisbasedonthedeviationfromcircularityto TheclassificationresultislistedinTable3and ellipticity. Further differences are reflected by illustrated in Figure 8 using the 1st and 2nd several features sensitive to the different canonicalvariables(CAN1,CAN2)withtheclass arrangement of bright fluorescent particles discriminationlines.Usingthree-classdiscrimi- (HETERO)insidethedarkernucleussurround- nationanalysis,itisclearthatthegrandmgcell Fig.4. Originalfluorescenceimage(gr)(left),mg(right). Reduction of Anabolic Signals 443 Fig. 5. Effect of different filters on discrimination of features (a) original fluorescence image, (b) light intensityimage,(c)fluorescentintensityFfromREDchannelwithfrequencydistribution,(R)REDwith frequencydistribution,(G)GREEN,and(B)BLUEintensitychannel.(Colorfigurecanbeviewedintheonline issue,whichisavailableatwww.interscience.wiley.com.) populations are significantlydifferent from one From ground (gr) to 1 g (in-flight) (Table 4): another.ThefirstcanonicalvariableCAN1alone The selected features HETERO, ELNENL, permits a nearly complete discrimination at RHA, and NR4 represent pure textural aboutCAN1¼0.8ofthemgcellsfromtheothers changes. (seeFig.8).Asequenceofcellimagesisdisplayed From 1 g (in-flight) to mg (Table 5a): The inFigure9usingthefirstcanonicalvariableas sequenceofselectedfeaturesstartswithP2A,a order criteria. Note the change of particle shapefactorshowingtheinfluenceofgravityon structurefromlefttoright. growth form, followed by texture features HLM3andRHUA. Two-Class Cases From ground (gr) to mg (Table 5b): For Theywereanalyzedtofindoutmoreaboutthe completenessthedirecttransitionfromground changes from one experimental condition to (GR)tomgisalsooutlined.Chosenfeaturesare anotherintermsofmostselectivefeatures. HETERO,RHUA,andRATIO.Thefirsttwoare pure textural and the last is a shape feature similartoP2A. The two-class discrimination analysis pro- videdacomparableresultwiththeheterogene- ity of nuclear texture (HETERO) showing the greatest difference between gr and mg (correct class94.7%).Asimilardifference,withalower correlationwasobservedfor1gandmg(correct class 86.4%). Furthermore, the 1 g cells have very similar morphology to the ground cells (correctclass78.7%).Inotherwords,theground and1gsamplesaremoresimilartoeachother than they are to mg cells. The morphological deviationofmgcellsmaybeduetotheinfluence of gravity on the arrangement of nuclear Fig. 6. Segmentation with mask M by threshold (red) and r particlesasaresultofalteredgeneexpression. manualcuttinglines(green).(Colorfigurecanbeviewedinthe Allotherselectedfeatures(HLM3,RHUA,NR2, online issue, which is available at www.interscience.wiley. com.) NC9) except the last are features of particle 444 Hughes-Fulford et al. Fig.7. Illustrationsoffeatureextractionforonenucleus. arrangement. RATIO reflects an additional mg. Since changes in autocrine, growth, cytos- shapeinfluence. keletal,andtranscriptionfactorsmaybeimpor- tant in osteoblast physiology in spaceflight DISCUSSION [Cubano and Lewis, 2000; Goodwin et al., 2000], we analyzed 24 osteoblast genes under Combatingspaceosteoporosisisakeyfactor mg, 1 g, and gr conditions. Of the genes with in planning long-range space missions of significantlyalteredexpression,therewerefour greaterthan6months.Inordertounderstand cell cycle related, one osteoblast specific, two underlyingcellularmechanismsofdisuseosteo- immediate early, and two growth factor genes porosis, osteoblasts were grown in mg and 1 g reducedatoneorbothofthetimepoints.Many on the Space Shuttle and were analyzed for genes had no significant change in message changes in gene expression and morphology level,theseincluded18S,H4histone,fibronec- after growth stimulation. We noted visual tin, b-actin, cyclophilin, and prostaglandin E differences in nuclear shape in cells growth in 2 receptor 1 (EP1). In contrast, there were nine genes that had a significant reduction in message induction in mg, these included cox-2, cpla2, PCNA, bcl2, bax, c-myc, TGFb, OC, and fgf-2 suggesting a change in anabolic response intheabsenceofgravity. Of particular interest was the reduction in cox-2 message in the presence of high levels of PGE . Cox-2 is a key enzyme in the mainte- 2 nanceofboneinresponsetostressandexercise [Forwood, 1996; Forwood et al., 1998]. More- over, cox-2 is a feed-forward enzyme, which is induced by its own product, PGE [Pilbeam 2 etal.,1995].Inmg,thisfeed-forwardregulation becomes uncoupled, with PGE levels signifi- 2 cantly increased in both the 1 g (P<0.05) and the mg (P<0.001) osteoblasts when compared to gr controls. The underlying mechanism for this uncoupling is currently under study using modeled mg and understanding the loss offeed-forwardinductionofcox-2atthecellular Fig.8. Three-classdiscriminationresultusingthe1stand2nd levelmayhelpusunderstandthelossofbonein canonicalvariableswithdiscriminationlinesbetweenrespective classes. astronauts.
Description: