Heterogeneity in hormone responses and
patterns of collagen synthesis in cloned dermal
fibroblasts.
S R Goldring, … , S M Krane, J H Korn
J Clin Invest.
1990;
85(3)
:798-803.
https://doi.org/10.1172/JCI114506
.
Fibroblasts cultured from normal human dermis are heterogeneous with respect to growth
kinetics, synthetic function, and morphologic features. There are many examples of clonal
heterogeneity in apparently homogeneous connective tissue cell populations, and it has
been suggested that selection of cell populations with particular phenotypic features is the
basis for the development of pathologic connective tissue changes in inflammatory
disorders. In these studies we report characterization of the pattern of matrix biosynthesis
and responses to hormones in cells cloned from normal human dermis. The results indicate
that cloned dermal fibroblasts are heterogeneous with respect to synthesis of collagens as
well as their responses to prostaglandin E2 and parathyroid hormone. Selective expansion
of clonal populations with unique patterns of matrix synthesis and cell surface receptors
could provide the basis for abnormal connective tissue remodeling in certain pathologic
states.
Research Article
Heterogeneity
in Hormone
Responses and
Patterns
of Collagen Synthesis
in Cloned Dermal
Fibroblasts
Steven R. Goldring, Mary L. Stephenson,Elaine Downie, StephenM. Krane,and Joseph H. Korn
Department ofMedicine, Harvard MedicalSchool, Boston,Massachusetts02115; MedicalServices(Arthritis Unit),NewEngland DeaconessHospital,Boston,Massachusetts02215;Massachusetts GeneralHospital,Boston, Massachusetts02114; and the
Division ofRheumaticDiseases, VeteransAdministrationMedicalCenter, University ofConnecticut
School ofMedicine, Newington, Connecticut 06111
Abstract
Fibroblasts cultured from normal human dermis are heteroge-neous with respect to growth kinetics, synthetic function, and morphologic features. There are many examples of clonal het-erogeneity in apparently homogeneous connective tissue cell populations, and it has been suggested that selection of cell
populationswithparticular phenotypic features is the basis for the development of pathologic connective tissue changes in
inflammatorydisorders. In these studies we report character-ization of the pattern of matrix biosynthesis and responses to hormones in cells cloned from normal human dermis. The re-sultsindicate that cloned dermal fibroblasts are heterogeneous with respect to synthesis of collagens as well as their responses
to prostaglandin E2 and parathyroid hormone. Selective
ex-pansionof clonal populations with unique patterns of matrix
synthesisand cell surface receptors could provide the basis for abnormal connective tissue remodeling in certain pathologic states.(J. Clin. Invest. 1990. 85:798-803.) hormone-collagen *connective tissue cell
Introduction
Examination of the growth kinetics and morphologic features
of fibroblastscultured fromneonatal human skin have dem-onstrated that these cells are genetically and phenotypically diverse(1-7). Examplesofclonalheterogeneity inapparently
homogeneousconnective tissuecell populations have been
re-ported, and it has beensuggested that temporal selection of individual cell populations could be involved in cellular
re-modeling of connective tissues under physiologic conditions
(1-9). Similar mechanisms could also be involved in
patho-logic states in which monoclonal or oligoclonal cell
popula-tionspossessing unfavorableorinappropriate phenotypic
fea-turescould beselected(10-15).
The initialgoalofthesestudieswas todeterminethedegree ofphenotypic diversity in the pattern of matrix biosynthesis
among normal skin fibroblasts. We observed significant dif-ferences in the pattern of collagen
synthesis
in substrainscloned from normal neonatal foreskin. Thesefindingscanbe
Address correspondence to Dr. Steven R. Goldring, Massachusetts GeneralHospital-East,149 TheNavyYard, 13th Street,Charlestown, MA02129.
Receivedforpublication25 July1989 andinrevisedform 9
No-vember1989.
accounted for in part, by differential regulation ofcollagen
synthesisatthetranscriptionallevel.
The second goal of our studies was to determine whether heterogeneity also exists in responses involving cell surface receptors for hormones known to affect connective tissue re-modeling. We therefore examined the pattern and magnitude
of3'-5'-cAMP responses to PGE2 and PTH. Marked heteroge-neity in the cAMP responses to these hormones was also
ob-served, although there was nocorrelationbetween the rate of collagen synthesis and the magnitude or pattern of hormone responses. Selective expansion of clonal populations in vivo could thus give riseto tissues withunique patterns of matrix
synthesisand altered responses to hormonalregulation.
Methods
Processingoftissue. Samplesofneonatalforeskin obtained from
rou-tine circumcisionweredigested with clostridial collagenase. Substrains
wereisolated fromprimary cultures bylimitingdilution cloning(1-4).
Once confluent, cultures ofthese clones were maintained in basal Eagle'smedium supplemented withFCS andcarried through multiple
passages aftertrypsinization.
Hormone responses. Cellswerepassaged into24-well trays(Costar, Cambridge, MA)at - 5 X 104 cells/well. cAMP responseswere
as-sessedby incubatingcellsfor 20 minat370C inbuffercontaining 1 mM 3-isobutyl-l-methylxanthine (IBMX)' (Aldrich Chemical Co.,
Milwaukee,WI) inthe absence orpresenceof eitherPGE2orPTH. cAMP contentwasdeterminedby radioimmunoassay
(Becton-Dick-inson &Co.,Paramus,NJ)(16,17). PGE2was agiftoftheUpjohn Co.,
Kalamazoo, MI. BovinePTH(- 3,000 U/mg) extracted from TCA precipitates of bovine parathyroid glandsandpurifiedbygelfiltration
and ionexchange chromatographyoncarboxymethyl-cellulosewas a
gift ofDr.Henry T. Keutmann(Massachusetts GeneralHospital, Bos-ton,MA).
Labeling of medium collagens.Cellswereplatedat2X 105cells/
dishin 3.5-cmdishesand 2 d laterincubatedfor 24 h with
L-[5-3H]-proline (30
,tCi/mmol)
in the presence of50jig/mlbeta-aminopro-prionitrile, 50
jg/ml
ascorbicacid,and2 mMglutamineintheabsence of serum. We havepreviouslyfoundthatundertheseconditionsme-diumcollagenaccountsfor>95%of totalcollagensynthesized.
Me-diumproteinswhich includedprocollagensandprocessed procollagen were analyzed bySDS-PAGE with orwithout reduction with 0.5%
#-mercaptoethanol
andfluorographyaspreviouslydescribed(18). La-beled collagens were further characterized by pepsinization at 4°Cfollowedby SDS-PAGEwithorwithoutdelayedreductionto distin-guishal(I)fromal(III) collagenchains(18, 19).
Analysis of collagen mRNAs. CytoplasmicRNA was extracted from cellsbythe methodof WhiteandBancroft(20).After
denatur-ationin 7.5%formaldehyde,dothybridizationwas
performed
asde-scribedpreviously usingcDNAprobeslabeledbynick translation(21,
22).ThefollowingcDNAswereused: Hf677,a 1,500-bpcDNA
en-1. Abbreviation usedin this paper: IBMX,
3-isobutyl-l-methylxan-thine.
798 S. R.Goldring,M. L.Stephenson,E.Downie,S.M.Krane, and J. H.Korn
J.Clin.Invest.
© TheAmericanSociety for Clinical Investigation,Inc.
0021-9738/90/03/0798/06 $2.00
coding partof the al(I)procollagen subunit (23),and Hfl131,a 1,500-bp cDNA encodingpartof the a2(I) procollagen subunit (24) provided byDr.F.Ramirezand Dr. D. J.Prockop. ThecDNAfor the
a1(III)procollagen,pHCIII-I,wasprovided byDr. R.G. Crystal and
Dr. M.Brantly,National Institutes of Health (NIH; 25). Specific
4ctivi-ties ofthelabeledprobeswere - 2.16, 1.77, and 1.45 X108 cpm/yg
DNAforthea1(I), a2(I), anda1(111) procollagen cDNAs, respectively.
For Northern hybridization, total RNA was extracted usingthe guanidine-thiocyanate procedure (21). TotalRNA (5 Ag)was dena-tured bytreatmentwith formaldehyde-formamideand
electropho-resed on0.8%agarosegelscontaining 0.22Mformaldehyde.TheRNA wastransferredtonitrocellulosepaperandhybridizedusing thecDNA
probes describedabove(21). The relativeintensity of the 28Sand18S ribosomal RNA, compared with background as an index of RNA breakdown,wasassessedusingethidium bromide andvisualized under ultravioletlight. Nicktranslation,prehybridization,andhybridization withthe procollagen cDNAs wereperformedasdescribed (21).
Results
We have previously shown that cells cultured from human dermis increase cAMP content in response to PTH orPGE2 (26, 27). To examine the pattern of hormone-induced cAMP responsesin dermal fibroblasts, 24 cell substrains cloned by
limiting dilution from three neonatal foreskin sampleswere
incubated with maximal stimulatory concentrations of PTH orPGE2. The three parent cultures of fibroblasts were
respon-sivetoboth hormones. Allsubstrains preparedfrom the parent cultures from two of thesamples (D.W. and J.R.)increased cAMP contentwhen testedwithPTH orPGE2, althoughthere were marked differences in the magnitude ofthe cAMP re-sponses.Althoughcellsfromthethirdparent culture(L.T.F.)
were responsive to PTH, none ofthe substrains increased
cAMPlevels in response to PTH (Table I). These substrains were, however,responsiveto
PGE2-Fig. 1 shows therelationship between PTH- and PGE-in-duced cAMP responses in clones fromsampleD.W. Although there was a correlation of the PTH and PGE2 responses in
substrains from parent culture D.W.(r= 0.77),there wasno
correlation in the responses to the two hormones in the sub-strains from parentculture J.R. (not shown). Patterns of
re-sponsivenesstoPTHandPGE2remained stableafterpassage; in general,substrainsexhibitingthe greatesthormone-induced increasesin cAMPlevelscontinuedtoshow enhanced sensitiv-itytohormonai stimulation.
TableL ComparisonofPTH-induced cAMP Response inaParentDermalCell Cultureand Two
RepresentativeSubstrains
cAMP
Parent Substrains
L.T.F. C K
pmol/well
Buffer 1.84±0.45 0.55±0.13 0.45±0.06 PTH 7.60±1.82 0.52±0.06 0.57±0.04
PGE2 71.1±1.84 47.7±4.03 6.40±0.93
Cellswereincubated for20 min at 37°C with buffer containing
(1 mM)IBMXwith or without PTH(l0-7M) or±PGE2(2.5 X 10-6 M). Valuesrepresentmean±SEM.
o)
0 U) Q 0 0.0).
L.=
0
so
U)
C00
2-tO
,
T-0
._0
la I.
0 20 40 60 8 0
PGE2-induced cAMP response (pmol/5X104cells)
Figure1.Comparison of thePTH- andPGE2-inducedcAMP
re-sponses to maximalstimulatory concentrations of hormone in
sub-strains fromparentdermal cultureD.W.Cellsin 24-welltrayswere
incubated 20 minat370 inthe presenceofImM IBMX±PTH(l0-7
M)orPGE2(2.5X 106 M). Pointsrepresent meansof threewells foreach hormone.
We have previously described an inverse relationship
be-tweenendogenous PGE2 production and the magnitude of the PGE2-induced cAMP response in different human connective tissue cell cultures (28-30). As shown in Fig. 2, therewas no
correlation between basal PGE2 levels and the cAMPresponse tothis ligand in the D.W. substrains.
To furtherinvestigate thefunctional heterogeneityof the
dermalcells,substrains from the threeseparate parentcultures were examined for their capacity to synthesize collagen. All substrains produced types I and III collagen (analyzed after pepsinization and resolution of the labeled medium proteins by SDS-PAGE andfluorography),although therewas variabil-ity in the amount and relative abundance of the individual collagentypes.To quantitate the relative capacity of substrains
toproduce individual procollagens, cells were incubated with [3H]proline, and after separation of the pepsin-resistant
me-0
U)-)0)
.=
2-O
0CD
0 -40
0E
-Q1
100
80
60
-40
-20
-O-_
0.0 0.5 1.0 1.5 2.0
PGE (10
8M)
2
Figure2.Comparison of the PGE2-induced cAMPresponseand
basalPGE2 productionin substrainscloned fromonedermal
sam-ple. Cells in 24-welltrayswereincubated 20 minat37°Cinthe
pres-ence orabsence ofamaximalstimulatoryconcentrationofPGE2 (2.5X10-6 M). PGE2 levelsweremeasured after72 h incubationof
substrains(5X 104cells)in culturemedium alone.Values represent
meansofthreewells.
S
0
B Figure 3. Capacity of
der-
0.-A.< 4000 mal substains to
incorpo-O,. 4000 |-Z@ ok 1 * rate[3H]proline into
._ =L aI(I) procollagen;
com-3000
*
Cf
anparison with basalPGE2E
3000-
production andPGE2-in-oE 0~~~ ~ ~ ~ ~ ~~~~~~0m 22000- ducedcAMPresponse.
-w MediumPGE2 was
as-g 2000 * sayed in substrains (5
.2 u
C0~~~~~~~~~~~~~~~~~;-0*
.27Co 00 X104
cells) incubated 240~~~~~~~~
aoe G2 epne000 * h in culture medium
o . 0 alone.
PGE2
responsesc _ 0 . were' t o-. ,assessed. ,bytest-in-
, .
, .
0 0 1 0 20 3 0 4 0 5 0 cubatingcells(5 x10
0.0 0.2 0.4 0.6 0.8 1.0 cells/well) for 20min
PG18M PGE2-induced cAMP response wt rwtotPE
PGE2(10
M)(pmol/well)
withorwithoutPGE2(2.5X 106M). A com-paresPGE2productionto a1(I) procollagensynthesis. Pepsin-resistent labeled medium proteins were separated by SDS-PAGE. The regions of
thegel containingthea1(I)procollagen bands were excised and counted. Counts have been normalized to DNA content in the culture dishes at thetimeof labeling.PGE2productionand a1(I) procollagensynthesis were negatively correlated (r = 0.33). B comparesa1(I)procollagen
syn-thesis withthePGE2-inducedcAMP responses, which were negatively correlated (r = 0.21).
dium proteinsbySDS-PAGE the labeled bands were excised and counted. Fig. 3 shows incorporation of [3H]proline into
a1(I) procollagen in substrains cloned from a single sample of dermis. There was marked heterogeneity in the capacity of the
substrainstosynthesizea1(I)procollagen. With passage of the cells the pattern ofcollagen synthesis among substrains
re-mainedstable. There was no correlation between the amounts
ofa (I)procollagensynthesizedandthe production of PGE2 (r = 0.33) (Fig. 3 A) or the magnitude of the PGE2-induced cAMP response(r=0.21) (Fig. 3 B).
To determine whether the heterogeneity in collagen
syn-thesis amongthe substrains was correlated with steady-state levelsofprocollagen mRNA, cytoplasmic RNA was extracted andhybridizedwith labeled cDNAs for types I and III procol-lagen. Northern blot analysis of total RNA demonstrated that therewas nocross-hybridizationamong the probes (data not
shown). The levelsof procollagen mRNAdeterminedby dot
hybridizationasafunction ofthe secreted labeled procollagen
estimated by [3H]proline incorporation are shown in Fig. 4. Dotblotswerequantitatedbyscanningwith a soft laser densi-tometer and analyzed with an Apple Ile computer using a
Zeineh Videophoresis II electrophoresis reporting program
(Biomed Instruments Inc., Fullerton, CA) as previously
re-ported (21).Therewas acorrelationbetween the steady-state levelsofa1(I)procollagen mRNA and incorporation of
[3H]-proline into al(I)collagen chains(r = 0.68). No correlation wasfound in the pattern of collagen synthesis and hormone
responsiveness.
ThelevelsofaI(I)anda2(I)mRNAswerequantitated in
substrains from one parent culture (L.T.F.) by scanningthe
dotblots after
hybridization
with cDNAprobes. The ratios ofthe a1(I) and
a2(I) procollagen
mRNAs inthe substrains ranged from 1.37 to 2.46. As shown inFig.
5, the ratios ofa
l(I)/a2(I)
mRNA wereinverselycorrelated with theratioof a1(III)/a
I(I) plusa2(I)
(type III/type
Icollagen).Comparison
ofthe levels oftypes I and III
procollagen
mRNA(Fig. 6)
revealedasignificant
positive
correlation(r
=0.86;
P<0.001).
Thecorrelationwasbest between the levels ofaI
(III)
anda2(I)
procollagen (r=0.89;P<
0.001)
compared
with that ofa (III)andal(I)(r= 0.79;P<0.01).
Discussion
We havepreviously demonstratedthat in responsetoproducts
present inmedium conditioned by PBMC, dermalfibroblast
populations cultured from different individuals exhibit
marked phenotypic diversity in proliferative responses and
synthesis ofPGE2 andcollagenase ( 1-4). This conditioned
me-dium containsmanybiologically active cytokines,suchas IL- 1 ortumor necrosis factor, that could
potentially
influencethefunctions of connective tissue cells. We speculated that the
heterogeneityinproliferative responsestotheseligandscould
provideamechanism for immune-mediated clonal selectionat
sites of inflammation by favoring overgrowth ofclonal cell
populations
withphenotypic
characteristicsthat promotedis-easeexpression. Ithas beenproposedthatseveralchronic
in-2000
2 1500
Z~~~~~
0 1000 2000 3000 4000
al(1)
protein(cpm/ggDNA)
Figure 4.Comparisonindermal substrains ofthea1(I)procollagen
mRNAlevelsandthe amountofsecreteda1(I)procollagen
esti-matedby[3H]prolineincorporation. After transfer of cytoplasmic
RNA tonitrocellulose and hybridization with the labeled procollagen
cDNAstheautoradiographswerequantitated by densitometric
scan-ning. Incorporation of[3H]prolineintoal(I)procollagenwas
deter-mined by counting excisedbandsfrom SDS-PAGE gels after
separa-tion oflabeledmediumproteinsasdescribedinFig. 3. Eachpoint
represents the valueforasingle substrain.
2 1
0'
0 2 4 6
al(III)/Total ca(I)+a2(I)
(Percent)
Figure5.Comparisonin dermal substrains of the ratio ofal(I)/a2(I) procollagenmRNAwith theratio oftypeIII/typeIprocollagen mRNA.After extractioncytoplasmicRNAwastransferredto nitro-cellulose andhybridizedwith32P-labeledaI(I),a2(I),or a
1(III)
pro-collagen probes.RNAwasquantitated by densitometricscanningof theautoradiographs.Eachpointrepresentsthevaluesforasingle
substrain.
flammatory
conditions associated with increased connective tissuedeposition
such ashepatic cirrhosis,
interstitialpneu-monitis, nephritis,
and scleroderma (7,31-33)
could be ex-plained by selectionin vivo of dominantclonesofcells thatproliferateandproduce
relatively
largeamountsofmatrix.Inthesestudieswefurtherevaluateddermalfibroblast
pop-ulations by measuring
theproduction
of extracellularmatrixmacromolecules by substrains derived from parent cultures. Marked differences in the pattern and amounts of
collagen
synthesis
weredemonstrated.Furthermore,
thedifferencesin the amountsand types of secretedprocollagens
estimated by[3H]proline
incorporation
correlated with the steady-statelevels of the
corresponding
procollagen mRNAs. Althoughthese resultsareconsistent with differential
regulation
ofcol-lagen synthesis at a
transcriptional
level, they could also be accounted forby
differences in mRNAstability.
Directmea-Z 100 E
75-0 E050 =-u 25
0*
0 500 1000 1500 2000
Type
Iprocollagen
mRNA(densitometry units)
Figure6.Comparison ofthe levelsoftype III [a
1(III)]
totype I[al(I)+a2(I)]procollagenmRNAinsubstrains fromparent culture L.T.F.After extraction cytoplasmicRNA was transferred to
nitrocel-luloseandhybridized with32P-labeleda(I), a2(I),oraI(111) procolla-genprobes. RNAwasquantitated by densitometric scanning ofthe
autoradiographs.Eachpointrepresents thevalues (expressed in den-sitometryunits)forasinglesubstrain.
surementsof
transcription
and mRNAstability
arerequired
toestablish the molecular basis for
heterogeneity
in matrixsyn-thetic
capacity
among substrains.We
previously
noted variation in the ratios of levels of a1(I)
anda2(I)
procollagen
mRNAs in cultured humanchon-drocytes
andsynovial
fibroblasts( 19, 21, 22, 34).
These resultssuggested
that theexpressionoftheseprocollagengenes isnotcoordinately regulated.
Wespeculated
that the excessa2(I)
procollagen
mRNAwaseithernottranslatedorthe translatedproduct
wasnotsecretedby
the cells. Similar variation in the ratio of the levels of theaI(I)
andac2(I)
procollagen mRNAswas observed among the dermal substrains in the present re-port, consistent with differential regulation ofthese two
pro-collagen
mRNAs.Furthermore,
the results would indicate that the normal ratio ofa1(I)/a2(I)
of 2:1 in parent cultures may reflect the average of constituentpopulationsthat varysignifi-cantly
from this ratio.Miskulinetal.
(25)
previously reportedthat theexpression
ofthe type III
collagen
gene is modulated coordinately with that of thetypeIcollagen
genes basedoncomparisonofpro-collagen mRNA levels in dividingand nondividing cells. In
ourstudieswedidnotobserveacorrelation between
collagen
synthesis
and celldoubling
times. However,the levels oftype
IIIprocollagen mRNAwerestronglycorrelated with those of type I(r=0.86;P<0.001).This correlationwasbest between
al(III)
and a2(I) (r = 0.89; P < 0.001)compared with that betweenalI(III)
andal(I) (r
=0.79;
P<0.01).Our resultsare consistent with coordinateregulation of types I and III colla-gen colla-geneexpressioninthese substrainpopulations.
Botsteinetal.(7)havereportedthat there is marked
heter-ogeneityin thecapacityof cloned human foreskin fibroblasts
toincorporate [3H]proline into collagens, although thespecific
collagen
types synthesized were not characterized. Morere-cently, Crouchetal. (35)have reported that
morphologically
distinct clones ofahumantumorcellline exhibit
heterogene-ityin theproductionof collagenaswellasfibronectin. Dermal fibroblasts cultured from patients with scleroderma
produce
more types I and III collagen than fibroblasts from normal
subjects (36-38);this enhanced collagen synthesis is accompa-niedbyincreased steady-state levels of types I and III procol-lagenmRNAs. Since the scleroderma phenotype persistsover
many passages in vitro these observations are consistent with the selection of populations of fibroblasts in vivo which are
characterizedby highratesof collagen synthesis. Our
findings
ofheterogeneityin the pattern of matrix synthesis among sub-strains derived from normal dermal fibroblasts providea ratio-nale for apparent clonal selection of fibroblasts in a disease
suchasscleroderma. Of interest in this regard are our earlier observations that short-term exposure of dermal cell cultures to mononuclear cellconditioned mediumresultsin selection of cell populations with enhanced proliferative and
PGE2-syn-thetic responses upon reexposure to this mononuclear cell conditioned medium (1). In analogous studies, Duncan and Berman have shown that transient exposure of dermal fibro-blaststo y-interferon leadsto apersistent decrease incollagen
synthesis (39). Inflammatory cytokines may thus have
long-lasting effects on the phenotype of connective tissue cells. In this report we have also described responses of fibroblast substrains to two hormones that affect connective tissue
re-modeling. Since PGE2 has been shown to modulatecollagen
there was acorrelation between PGE2 responsiveness and the level of collagen synthesis (22, 34, 40). We had previously shown that theresponsiveness of many different connective tissue cell types (including dermal-, synovial-, and
bone-de-rived cells) to PGE2 is markedly influenced by the levels of endogenous PGE2 (28-30). Treatment of cultured cells with mononuclearcell products such as IL-1/3 or tumor necrosis factor-a, which increased endogenous PGE2synthesis, resulted
in desensitization to PGE2 stimulation. These results were consistent withdownregulation of receptors by high ambient levels of PGE2. In the present studies we did not observe a correlation between endogenous levels ofPGE2 production andPGE2-induced cAMP responses. Basal unstimulated levels ofPGE2were very low (1.6-7.5 ng/ml) in medium from these cells, however, and based on previous studies with skin and
synovial fibroblasts (28, 29) higher levels of PGE2 may be required for receptor downregulation.
We had previously reported that cells cultured from human dermis increase cAMP levels after incubation with PTH(16, 26, 27).Wedemonstrated byphotoaffinity labeling
thatthePTHreceptor in these cells was similar to that in other PTHtargettissues (bone andkidney)(41). Others have
con-firmed our observations concerning PTH responsiveness in dermal cultures, although the kinetics of PTH binding may differ in skin- andbone-derived cells(42-44). We were inter-estedin examining the cells for responsestoPTH, since this hormone may also modulate collagen synthesis (possibly by inducing increasesinintracellularcAMPlevels) (45, 46).Our present resultsindicate that, similar to PGE2, the substrains exhibit markedheterogeneity intheirresponsivenessto PTH. Infibroblast substrainsobtained from one sample, PTHand PGE2 werepositively correlated, butthiswasnot observedin
substrains from anothersample. These results favor
heteroge-neityin cellsurfacereceptorfunctionratherthan ageneralized
up- ordownregulation ofadenylate cyclase activity. We have
not examined thesubstrains forthe capacitytoproducethe recently described PTH-related peptide (47-51). Thisligand
doesincreasecAMP levels inPTH-responsivetissues
includ-ing dermis, and could
potentially
downregulatePTH receptors in a manneranalogoustoendogenous PGE2.One factor
potentially
contributingtotheheterogeneityinthe patterns of collagen
synthesis
among the substrainscould
be related to differences ingrowth ratesand the effects of in
vitroaging (5, 6)oncellphenotype.This would berelevantto
the datain Table I, in which the parent (L.T.F.) was PTH
responsiveandsubstrainswereunresponsivetothis hormone. The parent cells hadundergonefar fewerdoublingsin
compar-ison
totheclones. However,thestability
ofsubstrainpheno-type with respect to collagen synthesis and PGE2 response
found in these studies and in
previous
investigations (1-3)arguesagainst
aging
or senescence asthecauseof thepheno-typic diversity.
Inconclusion,ourstudies demonstrate that connective
tis-suecell substrainsclonedfromnormal human dermal tissues areheterogeneousbothintermsof
synthetic products
and in theactivity
of cell surfacereceptorsforhormones thatregulate
connective tissuecellfunction. Selective
expansion
of cellpop-ulations of connective tissue cells in vivo could
give
risetotissues withaberrent patterns of matrix
synthesis.
Inaddition,
alterations in responsestohormones that
regulate
these cellscouldcontributetothedisturbance innormalfunctional
activ-ity.
Similarprocesses may also be involved inconnective tissueremodeling
underphysiologic
conditions.WethankMicheleAngelo for preparation of the manuscript.
This work wassupported by NIH grantsAM-03564, AR-32343,
and AR-20621,andgrantsfromtheMedicalResearch Service of the VeteransAdministration and the Arthritis Foundation.
This is publicationnumber1056ofthe Robert W. Lovett
Memo-rial Group inthe Study of Diseases Causing Deformities. References
1.Korn,J.H.,D.Torres, and E.Downie. 1983. Fibroblast
prosta-glandin E2 synthesis. Persistence ofan abnormal phenotype after
short-term exposure to mononuclear cell products. J. Clin. Invest.
71:1240-1246.
2.Korn, J. H.,D. Torres, and E.Downie. 1984.Clonal
heterogene-ity in the fibroblastresponse to mononuclear cellderivedmediators.
Arthritis Rheum.27:174-179.
3.Korn,J. H.,C.E.Brinckerhoff,andR.L. Edwards. 1985.
Syn-thesisof PGE2, collagenase and tissue factor by fibroblast substrains: substrainsaredifferentially activated for different metabolicproducts.
CollagenRelat.Res.5:437-447.
4.Korn,J. H. 1985. Substrain heterogeneity in prostaglandinE2
synthesis of human dermal fibroblasts: differences in prostaglandinE2
syntheticcapacity of substrainsare notstimulus-restricted. Arthritis Rheum. 28:315-322.
5.Smith,J.R., andL.Hayflick. 1974. Variation inthelife-span of clonesderived from human diploid cell strains.J.Cell Biol.62:48-53.
6. Martin, G. M., C. A. Sprague, T. H. Norwood, and W. R.
Pendergrass. 1974. Clonal selection, attenuation anddifferentiationin
aninvitromodel ofhyperplasia.Am.J.Pathol. 74:137-154.
7.Botstein, G. R., G.K.Sherer, andE.C.LeRoy. 1982. Fibroblast selection in scleroderma: an alternative modelof fibrosis. Arthritis Rheum. 25:189-195.
8. Ko, S. D., R.C. Page,and A.S. Narayanan. 1977. Fibroblast heterogeneity and prostaglandin regulation of subpopulations. Proc.
Natl.Acad. Sci. USA. 74:3429-3432.
9.Brinckeroff, C.E.,andJ. E.Nagel. 1981. Collagenase production by clonedpopulations of rabbit synovial fibroblasts.CollagenRelat. Res. 1:433-444.
10.Benditt,E.P., andJ. M.Benditt. 1973. Evidencefora monoclo-nal origin of human atherosclerotic plaques. Proc. Natl. Acad. Sci.
USA. 70:1753-1756.
11.Hassell,T.M.,R.C. Page,A.S. Narayanan, and C. G. Cooper. 1976. Diphenylhydantoin (Dilantin) gingivalhyperplasia:
drug-in-ducedabnormalityof connective tissue.Proc. Natl. Acad. Sci. USA. 73:2909-2912.
12.Liotta,L. A. 1986.Tumorinvasion andmetastases-role of the extracellular matrix: Rhoads Memorial AwardLecture.CancerRes.
46:1-7.
13. Liotta, L. A., C. N. Rao, and S. H. Barsky. 1983. Tumor
invasionand theextracellular matrix. Lab.Invest.49:636-649.
14. Whiteside, T. L., M. Ferrarini, P. Hebda,and R. B.
Buck-ingham. 1988. Heterogeneous syntheticphenotype of cloned
sclero-dermafibroblastsmay be duetoaberrantregulationin thesynthesisof connective tissues. Arthritis Rheum. 31:1221-1229.
15.Pearson,T.A.,J. M.Dillman,K.Solez,andR. H.Heptinstall.
1978. Clonalmarkers in thestudy oftheoriginandgrowthof human atherosclerotic lesions. Circ.Res.43:10-18.
16.Goldring,S. R.,J.-M.Dayer,R.G. G.Russell,H.J.Mankin, andS.M. Krane.1978. Responsetohormonesof cells cultured from humangiantcelltumorsofbone. J.Clin.Endocrinol. Metab.
46:425-433.
17.Goldring,S. R.,M.S.Roelke,K. K.Petrison,andA. K.Bhan. 1987.Humangiantcelltumorsofbone.Identification and character-ization of celltypes. J.Clin.Invest.79:483-491.
18.Amento,E.P.,A. K.Bhan,K.G.McCullagh, and S.M. Krane.
1985. Influences ofgammainterferononsynovialfibroblast-likecells.
Iainductionandinhibition ofcollagensynthesis. J. Clin.Invest. 76:837-848.
19.Stephenson, M. L., S. M. Krane, E. P. Amento, P. A. McCros-kery, and M. Byrne. 1985. Immune interferon inhibitscollagen
synthe-sisby rheumatoidsynovialcells associated with decreased levels of the
procollagen mRNAs. FEBS(Fed.Eur. Biochem. Soc.)Lett. 180:43-50.
20.White,B. A., and F. C. Bancroft. 1982.Cytoplasmicdot hybrid-ization:simpleanalysisof relative mRNA levels inmultiplesmall cell
ortissuesamples. J. Biol. Chem. 257:8569-8572.
21.Goldring, M.B., L.J.Sandell, M. L.Stephenson, and S. M. Krane. 1986. Immune interferon suppresses levels ofprocollagen
mRNAand type IIcollagen synthesisincultured human articular and costalchondrocytes.J.Biol. Chem. 261:9049-9056.
22.Goldring,M.B.,and S. M. Krane. 1987. Modulationby
recom-binant interleukin I ofsynthesisof types I and IIIcollagensand
asso-ciated procollagenmRNAlevelsincultured human cells.J.Cell. Biol. 262:16724-16729.
23. Chu, M.-L.,J. C. Myers, M. P. Bernard,J.-F. Ding,and F.
Ramirez. 1982. Cloning and characterization of five overlapping cDNA'sspecificfor the humanproa(I)collagenchain. Nucleic Acids
Res. 10:5925-5934.
24. Bernard, M. P., J.C. Myers, M.-L.Chu, F. Ramirez, E. F.
Eikenberry, and D.J. Prockop. 1983. Structure ofacDNA for the proa2 chain ofhuman type I procollagen: comparison withchick cDNA for proa2(I) identifies structurally conserved features of the
proteinand thegene.Biochemistry. 22:1139-1145.
25. Miskulin, M., R. Dalgleish, B. Kluve-Beckerman,S. I. Ren-nard, P.Tolstoshev,M.Brantly,and R.G.Crystal. 1986.Human type IIIcollagengeneexpression is coordinatelymodulated with the type I
collagen genesduring fibroblast growth.Biochemistry.25:1408-1413.
26.Goldring, S. R., J. E. Mahaffey, M. Rosenblatt,J.-M. Dayer, J.T. Potts,Jr., and S. M.Krane. 1979.Parathyroidhormone inhibi-tors:comparisonofbiological activityinbone- and skin-derived tissue. J.Clin. Endocrinol. Metab. 48:655-659.
27.Goldring, S. R.,M.S. Roelke, F. R.Bringhurst,and M.
Ro-senblatt. 1985.Differential effects ofparathyroidhormone responsive cultured human cellsonthebiological activityofparathyroidhormone and parathyroid hormone inhibitory analogues. Biochemistry. 24:513-518.
28.Goldring,S.R.,J.-M.Dayer,andS.M. Krane.1984. Rheuma-toidsynovialcell hormone responses modulated bycell-cell
interac-tions.Inflammation. 8:107-121.
29. Dayer,J.-M.,S.R.Goldring,D. R.Robinson,andS.M. Krane.
1979.Effects of humanmononuclear cellfactoroncultured rheuma-toidsynovialcells:interactions ofprostaglandin E2andcyclic
adeno-sine 3',5'-monophosphate. Biochim.Biophys.Acta.586:87-105. 30.Goldring,S.R.,M.S.Roelke, K. K.Petrison,A.E. Evins, and
S. M. Krane. 1987. Mechanisms by which monocyte-macrophage products regulate responsesofconnective tissue (bone and synovial)
cellstohormones.J. Bone Miner. Res. 2:239. (Abstr.)
31.Bitterman,P.B., M. D. Wewers, S.I.Rennard, S. Adelberg, and R.G. Crystal. 1986. Modulationofalveolarmacrophage-driven
fibro-blastproliferation byalternativemacrophage mediators. J. Clin.
In-vest. 77:700-708.
32. Wyler,D. J., S. Prakash, and P. Libby. 1987. Mesenchymal
target cellspecificityofegggranuloma-derived fibroblastgrowth factor
inschistosomiasis. J. Infect. Dis. 155:728-736.
33.Cotran,R. S. 1978.Monocytes, proliferation and
glomerulone-phritis. J. Lab. Clin. Med. 92:837-840.
34. Goldring, M. B., J. Birkhead, L. J. Sandell, T. Kimura, and S. M. Krane. 1988. Interleukin 1 suppresses expression of
cartilage-specifictypes II and IX collagens and increases types I and III collagens
in human chondrocytes. J. Clin. Invest. 82:2026-2037.
35. Crouch, E.C., K. R.Stone, M. Bloch, and R. W. McDivitt.
1987.Heterogeneity in the production of collagens and fibronectin by
morphologicallydistinct clones of a human tumor cell line: evidence
forintratumoral diversityin matrix protein biosynthesis. Cancer Res.
47:6086-6092.
36. Uitto, J., E. A. Bauer, and A. Z. Eisen. 1979. Scleroderma.
Increased biosynthesis of triple-helicaltype I and type IIIprocollagens associatedwithunaltered expressionofcollagenase byskinfibroblasts inculture. J.Clin.Invest. 64:921-930.
37.Graves, P.N., I. K.Weiss,J.S.Perlish,andR.Fleischmajer. 1983. Increased procollagenmRNAlevels in scleroderma skin fibro-blasts. J. Invest. Dermatol. 80:130-132.
38. Kahari, V. M., T. Vuorio, and K. Nasto-Salonen. 1984. In-creased type Icollagen mRNA levels in cultured scleroderma fibro-blasts.Biochim.Biophys.Acta. 781:183-186.
39. Duncan, M. R., and B. Berman. 1987. Persistence ofa
re-duced-collagen-producing phenotype in cultured scleroderma fibro-blasts after short-term exposure to interferons. J. Clin. Invest. 79:1318-1324.
40.Krane,S.M.,J.-M.Dayer,L. S. Simon, and M. S. Byrne. 1985. Mononuclearcell-conditioned medium containing mononuclear cell factor(MCF), homologouswithinterleukin 1,stimulates collagen and fibronectinsynthesis byadherentrheumatoidsynovial cells: effects of prostaglandin E2 and indomethacin. Collagen Relat. Res. 5:99-117.
41.Goldring, S. R., G. A. Tyler, S. M. Krane, J. T. Potts, Jr., and M. Rosenblatt. 1984. Photoaffinitylabeling of parathyroid hormone receptors:comparison of receptors across species and target tissues and afterdesensitization to hormone. Biochemistry. 23:498-502.
42. Pun, K. K., C. D. Arnaud, and R. A.Nissenson. 1988. Parathy-roidhormone receptors in human dermal fibroblasts: structural and functional characterization. J. Bone Miner. Res. 3:453-460.
43. Silve, C., A. Santora, N. Breslau, A. Moses,and A. Spiegel. 1986. Selective resistance to parathyroid hormone in cultured skin
fibroblastsfrompatientswithpseudo-hypoparathyroidismtypelb.J. Clin.Endocrinol. Metab. 62:640-644.
44.Silve, C., A.Santora,and A.Spiegel. 1985. A factor produced
bycultured rat leydig tumor (Rice-500) cells associated with humoral
hypercalcemiastimulates adenosine3':5'-monophosphateproduction via the parathyroid hormone receptor in human skin fibroblasts. J.
Clin. Endocrinol. Metab. 60:1144-1147.
45. Kream, B. E., D. W. Rowe, S. C. Cwarek, and L. G. Raisz. 1980. Parathyroid hormonealters collagen synthesisand procollagen mRNA levels in fetal rat calvaria. Proc. Natl. Acad. Sci. USA. 77:5654-5658.
46. Berg, R. A., J. Moss, B. J. Baum, and R. G. Crystal. 1981. Regulation of collagen production by
fl-adrenergic
system. J. Clin. Invest. 67:1457-1462.47.Suva,L.J., G.A. Winslow,R. E. H.Wettenhall,R.G. Ham-monds, J. M. Moseley, H. Diefenbach-Jagger, C. P. Rodda, B. E. Kemp, H.Rodrigues,E. Y.Chen, P. J.Hudson,T. J.Martin,andW. I. Wood. 1987. Aparathyroid hormone-related protein implicated in
malignant hypercalcemia: cloning and expression. Science (Wash.
DC).237:893-895.
48.Strewler, G.J., P. H.Stern,J.W.Jacobs,J.Eveloff,R. F.Klein,
S.C. Leung,M.Rosenblatt,and R. A.Nissenson. 1987.Parathyroid hormonelikeproteinfrom human renal carcinoma cells. J. Clin. In-vest. 80:1803-1807.
49. Stewart,A. F.,T. Wu, D.Goumas, W. J. Burtis, and A. E.
Broadus. 1987. N-terminal amino acidsequenceoftwonovel tumor-derivedadenylatecyclase-stimulating proteins: identification of para-thyroid hormone-like andparathyroid hormone-unlike domains.
Bio-chem.Biophys.Res.Commun. 146:672-678.
50.Nissenson,R.A.,D.Diep,and G. J.Strewler. 1988.Synthetic peptides comprising the amino-terminal sequence ofaparathyroid hormone-like protein fromhumanmalignancies:bindingto
parathy-roid hormonereceptors and activation of adenylate cyclasein bone cells andkidney.J.Biol. Chem.263:12866-12871.
51. Danks, J. A., P. R. Ebeling, J. Hayman, S. T. Chou, J. M.
Moseley,J.Dunlop, B. E. Kemp, and T. J. Martin. 1989. Parathyroid