-
REPORTS
Cyclosporin A Rapidly Inhibits Epidermal Cytokine Expression in
Psoriasis Lesions, But Not in Cytokine-Stimulated Cultured
Keratinocytes
James T. Elder,t Craig Hammerberg,* Kevin D. Cooper,t Takayuki
Kojima,* Rajan P. Nair,* Charles N. Ellis, t and John J. Voorhees*
'DepaILm.ent of Dermatology, University of Michigan, AIm Arbor; and
tDermatology Service, Veterans Affajrs Medical Center, Ann Arbor,
Michigan, U.S.A.
To better understand the cellular target(s) of cyclosporin
ac-tion in psoriasis, we have studied the effects of syste'mic
short-term (7 d), low-dose (3 - 7.5 mg/kg) cyclosporin A
administration on the expression of the cytokines interleukin
(IL)-8 and IL-1 fJ in psoriatic lesions. RNA blot hybridization
analysis of pretreatment keratome biopsies revealed. that
ex-pression ofbo~h cyt~ki~e mRNAs v.:as.highly variable fr?m
patient to patIent. Slglllficant covanatlOn of both cytokme rnRNA
levels was noted (r = 0.86, P < 0.0001). However, there was no
significant correlation between expression of either cytokine and
clinical severity, as measured by the pretreatment Psoriasis Area
and Severity Index (P ASI). IL-1 fJ protein levels measured by
enzyme-linked immunosorbent assay (ELISA) were highly correlated
with IL-1 fJ mRNA levels, indicating that the differences in
transcript levels ac-curately reflect differences in epidermal
cytokine protein.
The effectiveness of orally administered cyclosporin A (CsA) as
an antipsoriatic agent has been well docu-mented [1]. Because CsA
appears to be a relatively specific immunosuppressive agent [2],
its effective-ness in psoriasis has been cited as evidence for a
central
role of inunune system activation in the pathogenesis of this
disease [3,4]. Ho'Wever, at sufficiently high concentrations (2-10
Ilg/ml), CsA also has significant anti proliferative effects upon
cultured human and murine keratinocytes (reviewed in [5,6]), and
skin tissue CsA levels approach this range after oral
administration of effective antipsoriatic doses [7] . Therefore, in
addition to its immunosup-press ive actions, CsA might act directly
to inhibit the proliferation of psoriatic keratinocytes.
The psoriatic keratinocyte may also play an important role in
the initiation and/or maintenance of cutaneous inflammation in
psoria-sis through the production of chemotactic and
proinflammatory cytokines, adhesion molecules, and
antigen-presenting molecules such as HLA-DR (reviewed in [6,8]).
Because this phenomenon may in turn be elicited by exposure to
immune system-derived cyto-kines, such as interleukin (IL)-l,
interferon (IFN)-y, and tumor
Manuscript received February 22,1993; accepted for publication
June 30, 1993.
Reprint requests to: Dr. James T . Elder, Department of
Dermatology, University of Michigan Medical Center, C560 MSRB II,
Ann Arbor, MI 48109-0672.
Abbreviations: AMLR, allogeneic mixed leukocyte reaction; PASI,
Psori-asis Area and Severity Index; T AC, triamcinolone
acetonide.
Signific.ant reductio~s in both cytokine transcripts and in IL-1
(lunmunoreactive protein were noted in the high ex-pr~sslOn
subgroup after 1 week of cyclosporin A therapy, pnor to detectable
clinical improvement. In contrast to its pronounced effects on
epidermal cytokine expression in vivo and ~he allo~eneic mixed
lymphocyte reaction in vitro, cyclo-sponne A dId not inhibit the
induction of intercellular adhe-sion mol.ecule (ICAM)-l or IL-8
mRNAs by cultured kerati-nocytes In response to IL-1 fJ or the
combination of tumor necrosis factor (TNF)-a and interferon
(IFN)-y. These data suggest that .epidermal keratinocytes respond
to signals pro-duced by ~ctivated T cells by coordinate expression
of multi-ple cytokmes, and that cyclosporin A acts primarily
through b.lockade ofT cells,. rather than through keratinocyte
activa-tIOn. Key words: Immunology/skin disease/interleukins/
ICAM-1/chemokines.j Invest Dermatol1 01 :761-766, 1993
necrosis factor (TNF)-Cl', it becomes important to determine
~hether the effec~s of CsA on keratinocyte cytokine production in
111110 are exer~ed dIrectly upon the keratinocyte, or instead
reflect a blockade of Immune system activation.
CsA is an effective antipsoriatic agent at both high [1] and low
[7] doses, although the rate of clearing is dose dependent. We have
taken advantage of this observation to study the effects of
short-t~rm CsA tre~tme~lt on cytokine gene expression in keratome
biop-sIes of psorIatIc sk1l1 prior to detectable clinical
improvement. This study had two major objectives: i) to determine
whether reduced cytokine expression precedes clinical improvement,
thus providing sO.me eVIdence ~or the ~:u-ticipation of epidermal
cytokines in psori-aSIs pathogeneSIs, and u) to determine whether
the effects of CsA are exert.ed directly ?r indirectly on the
psoriatic keratinocyte. To accomphsh these objectives, we performed
Northern blot and en-zyme-linked. im~unosorbent assay (ELISA)
analysis of IL-1 p and IL-8 expressIOn 111 keratome biopsies before
and after 1 week of low-dose C~A treatment. In addition, we
compared the effects of CsA on psorIatIc leSIons in VitlO to its
effects on the induction ofIL-8 and intercellular adhesion molecule
(ICAM)-l mRNAs by IL-1, IFN-y, an.d T~F-Cl' in cult~red
keratinocytes. The results suggest ~hat a~phficatl?n of cytokme
signaling may occur via a pathway 1I1volv1l1g kerat1l10cyte-denved
IL-8 in psoriatic lesions ill VitlO. In addition, comparisou of ill
VitlO and ill vitro results suggests that CsA exerts one or more
indirect effect(s) upon the expression of kerati-nocyte-derived
cytokines in psoriasis, presumably by inhibiting the elaboratIon of
actlvat1l1g slgnal(s) by immune and/or inflammatory cells.
0022-202X/93/S06.00 Copyright © 1993 by The Society for
Investigative Dermatology, Inc.
761
-
762 ELDER ET AL
MATERIALS AND METHODS
Study Design and Tissue Procurement The cohort of 26 patients
stud-ied here was a subset of the 85 patients enrolled in a
multidose CsA trial conducted at the University of Michigan (7).
Patients received oral CsA at doses of3 mg/kg/d (nine patients), 5
mg/kg/d (12 patients), or 7.5 mg/kg/ d (one patient), or placebo
consisting of olive oil and polyoxyethylated oleic glycerides
(olive oil Labrafil base, four patients). Prior to and after 7 d of
treatment, lesional buttocks or thigh skin was subjected to
keratome biopsy at a depth of 0.2-0.4 mm using a Castroviejo
keratome. Normal volunteers and psoriatics not treated with CsA
were recrUited from the Southeast Mich-igan area and the Department
of Dermatology clinic population and biop-sied in an identical
fashion. All studies involving human subjects were ap-proved by the
Institutional Review Board of the University of Michigan, and
informed consent was obtained from each patient.
Cell Culture Primary cultures of normal human keratinocytes were
pre-pared as described [9) from keratome biopsies of adult
volunteers. Subcul-tures were expanded in keratinocyte growth
medium (KGM; Clonetics, San Diego, CAl, and used in the second to
fourth passage. Cells were pretreated for 2-16 h with CsA dissolved
in dimethylsulfoxide vehicle (0.1 % final concentration), with
triamcinolone acetonide (TAC) dissolved in 10% eth-anol vehicle
[10) (0.01 % final ethanol concentration) or with vehicle alone.
The cells were then treated with either IL-l cr. (5-100 ng/ml;
Dainippon, Osaka, Japan) or a combination ofTNF-cr. (20 ng/ml;
Amgen, Thousand Oaks, CAl and IFN-y (100 U/ml; Collaborative
Research, Bedford, MA). Cytokine stocks were prepared in sterile
phosphate-buffered saline (PBS) and stored in small aliquots at
-70'C prior to use. After 6 h of cyrokine treatment, cultures were
harvested for RNA isolation as described below.
RNA Isolation and Analysis RNA was prepared from snap-frozen
kera-tome biopsies by the guanidinium isothionate-cesium chloride
technique as previously described [11) , except that cesium
trifluoroacetate (Pharmacia, Piscataway, Nl) was used according to
the manufacturer's instructions. RNA was isolated from
keratinocytes using RNAzol (Tel-Test, Friends-wood, TX) as directed
by the manufacturer. Forty micrograms total kera-tome RNA, or 20
/-Ig total keratinocyte RNA (determined by optical density [00],60)
was electrophoretically separated on 1 % formaldehyde-agarose gels
and transferred to derivatized nylon membrane (Zeta-Probe; Bio-Rad,
Richmond, CAl. Blots were hybridized against 32P-labe led probes
prepared as described below, and were quantitated using either a
laser densitometer
. [11) or a phosphorimager [12) and normalized to cyclophilin as
previously described [11).
Plasmids and Hybridization Probes Plasmid DNAs were prepared by
alkaline lysis and precipitation in polyethylene glycol [13). cDNA
inserts were prepared by digestion with appropriate restriction
endonucleases fol -lowed by electrophoresis in
low-melting-temperature agarose gels. Inserts were 32P-labeled by
random priming, and had a specific activity of 1- 4 X 109 cpm//-Ig
DNA. The IL-l p (14), lipocortin II l11), and cyclophilin (11)
probes used in these experiments have been previously described.
The IL-8 probe was a 0.45-kilobase pair (kb) Eco Rl insert from the
plasmid pMDNCF 2-1 0.5 (15), the ICAM-l probe was a 1.9-kb XbaI
fragment containing the entire ICAM-l coding sequence derived from
the plasmid pGICAM-l [16), and [he cel lular retinoic acid binding
protein (CRABP)-Il probe was a 0.9-kb EcoRl fragment derived from a
human skin fibroblast library (12). The 36B4 probe was a 0.7 -kb
PstI fragment derived from p36B4 [17), which has recently been
shown to encode the human acidic ribosomal phosphoprotein PO
[18).
Data Analysis All keratome biopsy-derived RNA specimens were
blotted and sequentially rehybridized in parallel, allowing the
quantitative dara from a total of three blots to be pooled. IL-8,
IL-l p, and CRABP-Il densi-tometry or phosphorimager data were
normalized to cyclophil in to control for differences in loading
and RNA intactness. Statistical comparisons were made on the
normalized data using analysis of variance as previously re-ported
[6J.
ELISA Assay oflL-l cr. and p In a subset of patients, the week 0
and week 1 keratome biopsy specimens were divided, a portion being
used to prepare RNA and the remainder to assay IL-l immunoreactive
protein levels. Aqueous extracts of snap-frozen keratomes were
prepared using Dulbecco's PBS containing 0.03% polyethylene glycol,
and immunoreactive IL-1 pwas measured using a commercially
available kit (Cistron, Pine Brook Nj) as described (14).
Immunoreactive IL-l cr. was quantified using an ELISA com-posed of
monoclonal anti - IL-l cr. and polyclonal rabbit anti - IL-l cr.
(both from Dainippon, Osaka, japan) as described previously
[14).
Immunocytochemistry Portions of the same frozen keratome
biopsies used to prepare RNA were mounted in OCT compound (Miles
Diagnos-
~ 30 w (/) 25 +1
~ 20 w ~ 15 w ~ 10 ()
(/) 5 en
THE JOURNAL OF INVESTIGATIVE DERMATOLOGY
ct ~ O~---.---'----r---.---'---~--'---~
o 2 3 456 TREATMENT (WEEKS)
7 8
Figure 1. Clinical evaluation of study cohort. Closed circles,
CsA-treated (22 subjects). Opell circles, CsA vehicle (four
subjects). Asterisks, statistical signifi-cance at the 0.001 level
compared to week 0 (pretreatment) . The p value at week 1 for the
treated group was 0.1. This cohort is a subset of a previously
published study (7), and these data are reprinted by permission of
the N Engl J Med (324:277, 1991).
tics, Elkhart, IN) and 6-/-1 cryostat sections were prepared,
acetone-fixed, air-dried, blocked with 10% normal goat serum,
decorated with HLel MoAb (anti-CD45, 1 :20 dilution,
Becton-Dickinson, San jose, CA) , fol-lowed by fluorescein
isothiocyanate (FITC)-cOI~ugated affinity-purified goat anti-mouse
IgG (1:80 dilution; Organon-Teknika-Cappel, Durham, NC) , and
examined with a Nikon fluorescence microscope as previously
described [14J. The dermis and epidermis of slides stained with
either HLel or the IgGl kappa isotypecontrol MoAb MOPC 21 were
evaluated by three blinded observers on a scale of 0, 1, or 2 for
absent, moderate, or marked staining, respectively.
Allogeneic Mixed Leukocyte Reaction (AMLR) Lymphocytes were
isolated from peripheral blood of two healthy individuals by Ficoll
-Hypaque centrifugation. Ten milliliters of blood were diluted with
an equal volume of RPM! 1640 medium and layered over 15 ml
ofHistopaque-l077 (Sigma, St . Louis, MO) and centrifuged for 30
min at 500 X g. The lymphocyte band a[ the interphase was
collected, washed twice, and resuspended in RPM! 1640 containing
10% human AB serum at 5 X 106 cells/m!. The two-way mixed
lymphocyte reaction was initiated by mixing 100-/-11 aliquots of
the cell suspension from each of the two individuals (50,000 ce lls
each) in a round-bottom 96-well plate. The mixed cells were treated
with 1 /-II of a 200 X solution of CsA in DMSO, or a 1000 X
solution ofTAC in 10% ethanol. 90% sterile water [10) to obtain the
desired final concentration, or vehicle alone. Triplicate wells
were used for each concentration of drug. The cells were incubated
at 37'C in a humidified incubator under 5% CO2, On day 6 the cells
in each wel l were treated with 10 /lCi of lH-thymidine and
incu-bated for another 18 h. The ce lls were transferred to filters
using a cell harvester and washed with ethanol, and 3H-thymidine
incorporation was determined by scintillation counting.
RESULTS
Clinical Evaluation After 1 week of CsA therapy (3-7.5 mg/
kg/d), there was no significant difference in clinical severity
relative to baseline, as evidenced by the Psoriasis Area and
Severity Index (PAS I) [19]. However, with prolonged therapy,
significant clinical improvement w as seen in this cohort of
patients by 2 weeks (Fig 1, closed circles). In contrast, no
improvement was obtained in the pla-cebo group at any time point
(Fig 1, open circles).
Heterogeneity of Cytokine Expression In Vivo Figure 2 de-picts a
Northern blot comparison ofIL-8, CRABP-ll, and cyclophi-lin mRNA
levels in keratome biopsy specimens from normal and psoriatic
individuals. Note that IL-8 transcripts (approximately 1.8 kb) are
undetectable in normal epidermis , but are elevated in most
psoriatic les ions, albeit to variable extents. This variability
was not due to variations in mRNA loading or intactness, as
expression of cyclophilin was very similar in all samples, (Fig 2,
boltom pa"e~ . except for one normal sample and one psoriasis
sample that suffer RNA degradation during processing (lall e 7,
normal, and la lle 4, psoriasis) . Moreover, expression of CRABP-ll
transcripts ( - 1.4
I
-
YOLo 101, NO.6 DECEMBER 1993
NORMAL PSORIASIS 285-
185- ... .... IL-8
285-
185-......... CRABP-II
~ ••• I .... CYCLOPHILIN
Figure 2. Northern blot analysis of IL-B, CRABP-II, and
cyclophilin mRN A levels in keratome biopsies of normal and
psoriatic skin. Re-hybridi-zations of the same blot are shown. Each
lalle, RNA extracted from a differ-ent individual. Mobilities of
ribosomal RNAs are indicated to tlie lift. Por-tions of the results
shown in the middle and lower pallels have been published
previously [11] and are included here for comparison only.
kb) was increased to nearly the same extent in all the psoriatic
samples (Fig 2, middle panel).
As previously reported [14), low levels of IL-l p mRNA are
detectable by Northern blotting in some but not all psoriatic
lesions (Fig 3). In pretreatment specimens in wh.ich IL-8 mRNA I~ve
ls Were increased, IL-l p mRN A tended to be 111creased as well
(Fig 3) . The autoradiographic data from the 12 subjects shown in
Fig 3 and from an additional 14 subjects not shown were quantitated
by den-sitometry or phosphorimager and normalized to cyclophilin.
As expected on the basis of inspection of the blots shown in Fig 3,
pretreatment IL-l P and IL-8 mRNA levels were highly correlated (r
= 0.86, P < 0.0001, Fig 4).
Figure 4 also demonstrates quantitatively that pretreatment IL-l
Pand IL-8 mRNA levels were substantially variable from patient to
patient. To determine whether this variability in cytokine
expres-sion might be correlated with the degree of infiltration by
immune/ inflammatory cells, a semiquantitative analysis of HLel
immuno-fluorescent staining was performed in cryostat sections
prepared
285 -
185 -
5 5 3 3 3 3 5 5 0101010101010101 ~ .~t If· '11,' ...
,If''·\'''
Pt'
BLOT #1
2...1...2.2. CsA, mg/kg o 1 0 1 0 1 0 1 week
~ EtBr
1 ~IL-8
~ IL-1 f3
~CVCLO
.... ~ .... LlPO" BLOT #2
Figure 3. Northern blot analysis of CsA effects on cytokine mRNA
levels In psoriatic skin. Re-hybridizations of two different blots
are shown. Top /'llnel, EtBr, ethidium bromide-stained gel. Lift,
mobilities of ribosomal RNAs. In all other palleis, the
hybridization probe is indicated to the right. Cyclo, cyclophilin;
lipo II, lipocortin II .
CSA INHIBITS EPIDERMAL CYTOKlNES 763
100 • X r=0.86 c:( 80 P W ...J
-
764 ELDER ET AL
1000 IL-1 13 (n=21) IL-1 ('( (n=15)
~ 800 w .... 0 600 a: 11.
Ol 400 E C, Q.
200
0 BASELINE WEEK 1 BASELINE WEEK 1
Figure 6. Effects of CsA on IL-l a and IL-l p immunoreactive
protein (in pg IL-l per mg keratome extract protein) in psoriatic
skin. Week 0 versus week 1 p values: IL-l p, P = 0.004, IL-l a, p =
0.77. Error bars, SEM.
lated (r = 0.75, P = 0.0001, n = 22). Although IL-l a
transcripts are undetectable by Northern blotting under these
conditions [14], IL-t a protein was detectable, and pretreatment
IL-l a and P pro-tein levels also proved to be correlated (r =
0.52, P < 0.05, n = 15). However, CsA treatment had no effect on
IL-l a protein levels, whereas IL-l P protein, like IL-t P mRNA,
was significantly re-duced after 1 week (p < 0.004, Fig 6). CsA
Effects on Cytokine Expression in Cultured Keratinoc-ytes Figure 7
shows results representative of 12 experiments in which secondary
cultures of normal human keratinocytes were propagated in
serum-free KGM, pretreated either with CsA, TAC, or vehicle
control, then subjected to treatments with various cyto-kines. CsA
at doses up to 10 11M (12,000 ngjml) and TAC at doses up to 111M
(300 ng/ml) were completely ineffective in inhibiting the induction
of IL-8 mRNA in response to IL-t a or IFN-y plus TNF-a (Fig 7, and
additional experiments not shown). Variation of 'the time of CsA
pretreatment from 2-24 h had no effect on the results (data not
shown). As previously reported [20-22], ICAM-l mRNA was induced
markedly by IFN-y plus TNF in keratinocytes grown in KGM. However,
pretreatment with either CsA or TAC also had no effect on these
responses (Fig 7). IL-l treatment at 100 U jml produced only a low
and variable ICAM-! mRNA response, which appeared to be potentiated
slightly by CsA (Fig 7 and data not
285-
185-
NONE CSA 1 CSA 10 TAC 0.1 TAC 1 [PRETREA 1] l!"M) ,....-----,
r---w ,....-----, ,....-----, ,....-----, 1 2 3 1 2 3 1 2 31 231 2
3 TREATMENT
~ ICAM-1
~ IL-8
~36B4
Figure 7. Effects of CsA and TAC on cytokine-stimulated IL-S and
ICAM-l mRNA levels in cultured human keratinocytes.
Re-hybridizations of the same blot are shown. Upper pauei, mixture
ofICAM-l and lL-S probes of comparable specinc and total activity.
Lower panel, cyclophilin probe. Mobilities of ribosomal RNAs are
indicated to the left, and probes to the right. Pretreatment with
CsA and TAC was for 16 h prior to 6 h of treat-ment with the
agonists indicated by uumber above tl.ejigure. I, no treatment; 2,
IL-l a (100 ng/ml); 3, IFN-y (100 U/ml) plus TNF-a (20 ng/ml).
A z o ~ 120
a: 100 o Q.~
~ c5 80 oa: ~ ~ 60 wo z 0 40 ccf!. :E ~ 20 >-J: .... 0 I
C2...
B z o ~ 120 < ~ 100 Il.~
~
-
VOL. 101. NO.6 DECEMBER 1993
bystander," or in fact plays an active role in the initiation
and/or maintenance of the psoriatic lesion.
Keratinocytes may participate in the pathogenesis of psoriasis
through the generation of chemotactic signals for
immune/inflam-matory cells [6,8]. IL-8 is one candidate for such a
signal, as it is known to be chemotactic for neutrophils ill vivo
[27,28] and for lymphocytes ill vitro [29], and IL-8 RNA and
protein are increased in psoriatic lesions ([.30 - 32],. Fig 2).
I.L-.8 is ~ member of.the int~rcrine a family of cytokmes, wluch
are dlstmgUlshed by their localIzation on chromosome 4q21 [33,34] .
These cytokines are often considered as secondary response elements
of the cytokine cascade, because several of them share the property
of inducibility by other (primary) cytokines (namely IL-1 and
TNF-a) in a variety of cell types [33,34]. IL-8 also has the
properties of a secondary cytokine in cultured keratinocytes, as
its expression is induced by IL-1 [21] or by TNF-C\' acting in
concert with IFN-y [20]. .
In studying the effect of CsA on epidermal cytokine expression
in psoriasis lesions, we analyzed keratome biopsies after 1 week of
therapy, when no significant improvement in clinical status could
be detected. This time interval was selected to avoid the
confound-ing effect of overall reduction in lesion severity on the
results. In this study, IL-8 and IL-1 p were variably but
coordinately expresse~ in pretreatment biopsies (Fig 3), resulting
in a significant correlatIOn between pretreatment IL-1 p and IL-8
mRNA levels (Fig 4). Curiously, neither differences in clinical
status nor in the total num-ber of bone marrow-derived inflammatory
cells in the lesions could explain the observed variability in
cytokine mRNA levels. After 1 week, both cytokine transcripts were
markedly reduced (81- 85%), whereas CRABP-II transcripts were
reduced only slightly (8%) (Fig 5). Taken together with the high
correlation in pretreatment IL-1 p and IL-8 values, their
coordinate reductions early in CsA therapy strongly suggest that
both cytokine responses are affected by a com-mon proximal
stimulus. which is inhibited by CsA. .
It is worth noting that all patients, and not just those with
high levels of IL-8/IL-1 p expression, rapidly improved with CsA
ther-apy. This suggests ~ith~r that CsA.may inhibit distinct signal
trans-duction pathways m different patients, or that the IL-8/IL-1
p re-sponse is variable in space and/or time and is therefor~ not
always detected. We favor the latter explanation, because the
differences 111 pretreatment cytokine mRNA levels between patients
could n~t be accounted for in terms of the total numbers of bone
marrow-denved cells in the biopsy specimen. In further support of
this concept, one of four placebo-treated patients showed a large
increase in IL-8 and IL-l p mRNA between week 0 and week 1 (data
not shown).
Given the relative specificity of CsA for T cells [2], the
simplest explanation for these resul ts is that CsA inhibits the
ability of acti-vated T cells to produce signals leading to
increased keratinocyte cytokine expression, but does not directly
inhibit the ability ofkera-tinocytes to respond to such signals. To
test this hypothesis, we studied the effects of CsA on IL-8
expression by cytokine-stimu-lated cultured keratinocytes. We
focused on IL-8, rather than IL-l p, because the high levels ofIL-8
detectable in some lesions indicate that the bulk of the IL-8
expression in psoriatic lesions is occurring in keratinocytes.
Although IL-8 can be expressed in a variety of cell types found in
the skin, including T cells [35], macrophages [3~], and fibroblasts
[21], its expression at high levels by keratinocytes 111 psoriasis
has been convincingly demonstrated [32]. In contrast, IL-1 p mRN A
levels in psoriatic lesions are near the lo.wer limi~ of de
tec-rability of the Northern blot assay [14], which IS
approximately 1 copy per cell under the conditions used [37]. Thus,
it is unclear whether the IL-1 p mRNA detected by Northern blotting
in psoria-sis lesions is keratinocyte derived. CsA was unable to
inhibit cyto-kine-stimulated IL-8 expression in cultured
keratinocytes after ei-ther CsA or TAC treatment, at concentrations
far higher than those required to inhibit the l1~ixed lymphocyte
response (Figs 7, 8) .. This result strongly reinforces the
conclUSIOn that the keratmocyte IS not the direct target of CsA
action in psoriasis. The inactivity of CsA in this regard was not
limited to the IL-8 response, as the ICAM-1 mRNA response to IFN-y
plus TNF-a was unaffected, and the limited ICAM-1 response to IL-l
was slightly potentiated (Fig 7).
CSA INHIBITS EPIDERMAL CYTOKINES 765
These results are consistent with previous studies that show
that at concentrations as high as 1- 5 ,ug/ml, CsA does not
inhibit, and in fact may slightly potentiate, keratinocyte
expression oOL-I, GM-CSF, IP-10, and transforming growth factor
(TGF)-a [38,39] .
Regardmg the nature of the T-cell-derived signal(s), we
specu-late that limited expression of primary cytokine(s) either
directly by T cells or by infiltrating macrophages and/or dendritic
cells under the control of T cells might trigger the expression of
IL-8 at much high.er levels. in nearby keratinocytes. This
interpretation would be consl~tent With. the demons~ratlOn of
clusters of IL-8 - expressing kerat1110cytes directly overlymg
collections of leukocytes [32]. It is attractive to speculate that
one of these primary cytokines might be IL-1 P; however, our
previous studies of IL-l in psoriasis indicated that IL-l P
extracted from psoriatic lesions lacked biologic activity [14].
Among many other possibilities, T -cell-derived IL-8 itself could
be one of the signals linking T cells and keratinocytes. Thus, CsA
ha~ been ~hown to inhibit the expression of IL-8 and several
related Intercnne-a gene family members in T lymphocytes [35] but
not In lfolonocytes [35,36] . Alternatively, the relevant signal
for the productIOn of IL-8 in keratinocytes could be actual contact
with T cells or macrophages themselves [40]. Preliminary reports
have suggested that IL-8 can stimulate keratinocyte growth in
organoty-pIC ~ulture systems,' and limited effects of IL-8 on
keratinocyte proltferatlOn have been observed under serum-free
conditions [41]. ~hether IL-~ and related cytokines such as gro-at
directly con-tnbute to psonatlc epidermal hyperplasia in villa
remains to be deter-mined.
.The strong positive correlation between IL-1 P mRNA and pro-tem
levels before and after CsA therapy indicates that the low level of
expressIOn detectable by Northern blotting is reflective of the
a~ount o~IL~ 1 P actually present in psoriatic lesions. However,
the biologiC slgmficance of these observations remains obscure, and
review of the I~terature reveals many apparent inconsistencies in
the le~els of cytokme mRNAs, proteins, and biologic activities in
psori-atic leSIOns [14,~2,36,42-45]. Potential explanations for
these dis-parate observations include: i) translation of unstable
or rare mRNAs mto stable proteins; ii) antibody cross-reactions
with other keratinocyte proteins; iii) differential biologic
activities of intra- and extr~cellul~ ~ytokine pools; iv) the
presence of specific and non-speCific mhlbltory prteins; and v) the
uptake of cytokines synthe-Sized elsewhere by epidermal
keratinocytes. Experimental evidence suppo.rts th.e conc.ept that
the skin may serve as a depot for systemic cytokmes, mcludmg TNF-a
[46] and IL-6 [47] . Moreover, proteins as large as bov111e serum
albumin (molecular weight 69,000) can enter the epldermal
.compartment from the circulation [48]. Addi-tional careful
investigations will be required to fully elucidate the cellular
source(s) of epidermal cytokines ill villa.
Taken together with results demonstrating the coordinate
varia-bility ~n all three forms of MGSA/ gro in psoriatic lesions
and their :eductlOn after 1 ",:,eek of CsA treatment, t the results
presented here ll1dlcate that multiple keratmocyte-derived
cytokines are under co-ordinate regulation in psorias is. perhaps
by virtue of a shared signal transduction pathway.capable of
responding to T-cell-derived sig-nals. The reduction m ~L-8/gro
chemotactic activity after CsA treatme.nt would beyredlcted to
suppress further influx ofT cells and antlgen-presentmg cells
[6,8], contributing to eventual resolu-tion of the lesion. However.
substantial clinical resolution required 6-8 weeks of therapy (Fig
1), whereas cytokine expression levels were mar~edly reduced afte~
only 1 week (Figs 3, 5). It is possible that once 111ltlated, the
keratll10cyte may be capable of sustaining a hyperproiIferauve
response via CsA-insensitive autocrine mecha-nisms, such as
overexpression of the EGF-like growth factors
• Reusch MK. Studtmann M. Schroeder J-M. Stichcrling M.
Chris-tophers E: NAP / interleukin 8 is a potent mitogcn for human
kcratinocytcs ill lIitra (abstr). J Illvest Dermatal 95:460,
1990
t Kojima T. Cromie MA. Fisher GJ. Voorhees JJ. Elder JT: Gro-a
mRNA is selectively overexpressed in psoriatic epidermis and is
reduced by cyclosporin A ill lIillO, but not in cultured
keratinocytes. J hlVesr Dermarol 101 :767 -772. 1993 (this
issue)
-
766 ELDER ET AL
TGF-a and amphiregulin [49-52]. In any case, these results
strongly support the concept that esA acts rapidly but indirectly
~o reduce expression of IL-8 and IL-l P expressIOn by th.e ps~natlc
keratinocyte in lIillo, and suggest a role for these cytokmes m the
pathogenesis of psoriasis.
We tllank Drs. C. Larsen and K. Matsushima for t.heIL-8 cDNA
probe, Dr. Atlders Astrom for tile CRABP-II eDNA probe, Dr. R.
Blake PepitlSky fo r the lipocor/itl II probe, and Dr. S. W.
Caughmatl for the ICAM-1 cDNA probe. Ti,e expert statisti-cal
assistance oJTed Hamiltotl, M.S ., atld the skilled /echnical
assistance oJJi CIIOW, Diatle Boman, and Qiotlg Yatlg are gra
tefully acknowledged. We aregratifrll to Dr. JotJa/hatl Barker for
many hellif"l conversations atld acktlowledge his participation in
early experiments on 1L-8 expressiotl in psoriat.ic lesions.
Tllis work was slIpported itl part by tlte Departmwt of Veterans
Affairs UTE, KDC), USPHS award R29 AR 40016 aTE), tile Getteral
Clinical Research Cettler at tlte Utliversity of Michigatl
(M01RR0042 from the Natiotlal Center for Research ResollrCes,
National Itls/illl/es of Health) , atld tlte Babcock Memorial
Trust.
REFERENCES
1. Ellis CN, Gorsulowsky DC, Hamilton TA, et 01: Cyclosporine
improves psoriasis in a double-blind study. JAMA 256:3110 - 3116,
1986
2. Kahan BD: Cyclosporin A: A selective anti-T cell agent. C/i"
Hoemoto/ll:743-761, 1982
3. Cooper KD: Psoriasis: leukocytes and cytokines. Dermotol C/i"
8:737 - 745,1990 4. Barker JNWN: The pathophysiology of psoriasis.
La"ret 338:227 - 230, 1991 5. Kanitakis J, Thivolet J:
Cyclosporine: an immunosuppressant affecting epithelial
cell proliferation. Arch Dermoto/126:369-375, 1990 6. Wong R,
Winslow C, Cooper KD: Mechanisms of cyclosporin action in
psoria-
sis. [mmlltlol Today 14:69-74, 1993 7. Ellis CN, Fradin MS,
MessanaJM, BrownMD, Siegel MT, Hartley AH, Rocher
LL, Wheeler S, Hamilton TA, Parish TG, Ellis-Madu M, Duell E,
Annesley TM, Cooper KD, Voorhees 11: Cyclosporine for plaque-type
psoriasis. Results of a multidose, double-blind trial. N E"gl J Med
324:277 -284, 1991
8. NickoloffB: The cycokine network in psoriasis. Arch
Dermato/127:871, 1991 9. Elder JT, Fisher GJ, Zhang Q-Y, Eisen D,
Krust A, Kastner P, Chambon P,
Voorhees JJ: Retinoic acid receptor gene expression in human
skin. J ltlVest Dermato/96:425 - 433,1991
10. Marcelo CL, Tomich J: Cyclic AMP, glucocorticoid, and
retinoic modulation of illllitro keratinocyte growth.J [lIlIest
Dermato/81 :64S-68S, 1983
11. Elder JT, Tavakkol A, Klein SB, Zeigler ME, Wicha M,
Voorhees JJ: Protoonco-gene expression in normal and psoriatic
skin.J [lIlIest DermotoI94:19-25, 1990
12. Elder JT, Astrom A, Pettersson V, Tavakkol A, Griffiths CEM,
Krust A, Kastner P, Chambon P, Voorhees JJ: Differential regulation
of retinoic acid receptors and binding proteins in human skin.]
ltlVest DermatoI98:673-679, 1992
13. Birnboim HC, Doly J: A rapid alkaline extraction procedure
for screening recom-binant plasmid DNA. N"c/ Acids Res
7:1513-1523,1979
14. Cooper KD, Hammerberg C, Baadsgaard 0, Elder JT~ ChanLS,
Sauder ON, Voorhees 11, Fisher G: IL-l activity IS reduced 111
psotlatlc sk1l1 decreased IL-l and increased nonfunctionallL-l p.J
[mllll",0/144:4593-4603, 1990
15. Matsushima K, Morishita K, Yoshimura T, Lavu S, Kobayashi Y,
Lew W, Appella E, Kung H, Leonard E, Oppenheim J: Molecular cloning
of a human monocyte-derived neutrophil chemotactic factor (MDNCF)
and the induction of MDNCF mRNA by interleukin 1 and tumor necrosis
factor. J Exp Med 167:1883 - 1893,1988
16. Simmons D, Makgoba MW, Seed B: ICAM, an adhesion ligand for
LFA-l, is homologous to the neural cell adhesion molecule NCAM.
Natllre 331:624-627,1988
17. Masiakowski P, Breathnach R, BlochJ, Gannon R, Krust A,
Chambon P: Clon-ing of cDNA sequences of hormone-regulated genes
from the MCF-7 human breast cancer cell line. Nllc/ Acids Res
10:7895-7903, 1982
18. Laborda J: 36B4 cDNA used as an estradiol-independent mRNA
control is the cDNA for human acidic ribosomal phosphoprotein PO.
Nllc/ Acids Res 19:3998,1991
19. Fredriksson T, Pettersson V: Severe psoriasis-oral therapy
with a new retinoid. Determatologico 157:238-244, 1978
20. Barker JNWN, Sarma V, Mitra RS, Dixit VM, Nickoloff BJ:
Marked synergism between tumor necrosis factor-a and interferon-y
in regulation of keratino-cyte-derived adhesion molecules and
chemotactic factors.J Clill lolllest 85:605-608,1990
21 . Larsen CG, Anderson AO, Oppenheim 11, Matsush ima K:
Production of inter-leukin-8 by human dermal fibroblasts and
keratinocytes in response to interleu-kin-lor tumour necrosis
factor. [mnllwol 68:31- 36, 1989
22. Norris DA, Lyons MB, Middleton MH, Yohn11, Kashihara-Sawami
M: Vltravi-olet radiation can either suppress or induce expression
of intercellular adhesion molecule 1 (ICAM-1) on the surface of
cultured human keratinocytes.] [IIlIest DermatoI95:132-138,1990
23. Henseler T, Christopher E: Psoriasis of early and late
onset: Characterization of two types of psoriasis vulgaris.J Am
Acad Dermato/13:450- 456, 1985
THE JOVRNAL OF INVESTIGATIVE DERMATOLOGY
24. Friedmann PS, Ford G, Ross J, Diffey BL: Reappearance of
epidermal Langcrhans cells after PUVA therapy . Br J
Dermato/l09:301 -307, 1983
25. Morhenn VB, Orenberg EK, Kaplan J, pfendt E, Tere ll C,
Engleman EG: Inhibi-tion of Langerhans cell-mediated immune
response by treatment modalities useful in psoriasis. J I" vcst
Dermato/ 81:23 - 27, 1983
26. Cooper KD, Oberhelman L, Hamilton TA, Baadsgaard 0, Terhune
M, leVee G, Anderson T , Koren H: UV exposure reduces
im11lUnization rates and promotes tolerance to cpicutancous
antigens in humans: relationship to dose, CDla-DR+ epidermal
macrophage induction, and Langerhans cell depletion. Proc Notl Acad
Sci USA 89:8497 - 8501, 1992
27. Leonard EJ, Yoshimura T, T anaka S, Raffeld M: Neutrophil
recruitment by intradermally injected Ilcutrophil
attractant/activation protein-I. J [lIlIcst D
er-,"010196:690-694,1991
28. Swensson 0, Schubert C, C hristophers E, Schroder J-M:
Inflammatory properties of neutrophil-activating
protein-1/interleukin 8 (NAP-I / 1L-8) in human skin: a light- and
electron microscopic study.J [IIlIest Dermoto/96:682-689, 1991
29. Larsen CG, Anderson AO, Appella E, OppenheimJJ, Matsushima
K: The neu-trophil-activating protein (NAP-I) is also chemotactic
forT Iypbocytes. Sciell (e 243:1464-1466,1989
30. Schroder J, Christophers E: Identification of C5a des arg
and an anionic nentro-phil activating peptide (ANAP) in psoriatic
scales.J lolliest Dennato/87:53- 58, 1986
31. Nickoloff BJ, Karabin GD, Barker JNWN, Griffiths CEM, Sarma
V, Mitra RS , Elder JT, Kunkel SL, Dixit VM: Cellular localization
of interleukin-8 and its inducer, tumor necrosis factor-a in
psoriasis. Am J PotllO/138:129 - 140, 1991
32. Gillitzer R, Berger R, Mielke V, Muller C, WolffK. Stingl G:
Vpper keratino_ cytes of psoriatic skin lesions express high levels
ofNAP-l/IL-8 mRNA in situ. J [lIlIcst DemuttoI97:73 -79, 1991
33. Oppenheim JJ , Zachariae COC, Mukaida N, Matsushima K:
Properties of the novel proinflammatory supergene "intercrine"
cytokine family. A IIIIII Rev [m-mlllloI9:617 -648,1991
34. Stoeckle MY, Barker KA: Two burgeoning families of platelet
factor 4-related proteins: mediators of the inflammatory response.
New Biologist 2:313-323, 1990
35. Zipfel PF, Bialonski A, Skerka C: Induction of members of
the IL-8/NAP-l gene family in human T lymphocytes is suppressed by
cydosporin A. Bioch"", Biophys Res Comllll'" 181:179-183, 1991
36. Mrowietz V , Sticherling M, Mielke V, Schroder JM, C
hristophers E: Neutro-phil-activating peptide l/interleukin 8 mRNA
expression and protein secre-tion by human monocytes: effect of
cyclosporin A. Cytokill' 3:322-326, 1991
37. Elder JT, Sartor CI, Boman DK, Benrazavi S, Fisher GJ,
Pittelkow MR: Interleu-kin-6 in psoriasis: expression and
mitogen.iciry studies. A rc/' Denn R", 284:324-332,1992
38. Gallo RL, Grabbe S, Choi SS, Bleicher P, Granstein RD:
Cyclosporin increases granulocyte/macrophage colony-stimulating
factor (GM-CSF) activity and gene expression in murine
keratinocytes.J Itwest DermotoI98:274-278, 1992
39. Khandke L, Krane J , Ashinolt"R, Staiano-Coico L,
Granelli-Piperno A, Luster A, Carter DM, Krueger J, Gottlieb A:
Cyclosporine in psoriasis treatment: inhibi-tion of keratinocyte
cell-cycle progression in Gl independent of effects On transforming
growth factor-a/epidennal growth factor receptor pathways. Arcll
Dermato /1 27:1172- 1179, 1991
40. StoofTJ, Mitra RS, Sarma V, Dixit VM, Nickoloff BJ:
Keratinocyte activation following T-Iymphocyte binding.J hlllesl
Dermoto/98:92-95, 1992
41. Tuschil A, Lam C, Haslberger A, Lindley I: Interleukin-8
stimulates calciUlU transients and promotes epidermal cel l
proliferation. J Itwest DermatoI99:294-298,1992
42. Sticherling M, BOnlscheuer E, Schroder J-M, C hristophers E:
Localization of neutrophil-activating peptide-l/interleukin-8 -
immunoreactivity in nonnal and psoriatic skin.J Itlllest
Dermato/96:26 -30, 1991
43. Anttila HSI , Reitamo S, Erkko P, Ceska M, Moser B,
Baggiolini M: Interleukin-8 immunoreactivity in the skin of healthy
subjects and patients with palmoplan-tar pustulosis and psoriasis.
J Itwest Derllloto/98:96-101, 1.992
44. Gearing AJH, Fincham NJ , Bird CR, Wadhwa M, Meager A,
Cartwright JE, Camp RDR: Cytokines in skin lesions of psoriasis.
Cytokille 2:68 - 75, 1990
45. Prens EP, Benne K, van Damme J , Bakkus M, Brakel K, Bemler
R, von Joost T: Interleukin-l and interleukin-6 in psoriasis. J
[lIlIest Dermato/95:121S-124S, 1990
46. Beutler BA, Milsark IW, Cerami A: Cachectin/tumor necrosis
factor: produc-tion, distribution, and metabolic fate illllillo.J
ItIllIlIlIlO/135:3972-3977, 1985
47. CastellJ, KlapprothJ, Gross V, Walter E, Andus T, Snyers L,
ContentJ, Heinrich RC: Fate of interleukin-6 in the rat:
involvement of skin in its catabolism. EurJ
BiocllelllI89:113-118,1990
48. Rabilloud T, Asselineau 0, Dannon M: Presence of serum
albumin in norn:L1l human epidermis: possible implications for the
nutrition and physiology of stratified epithelia. Mol Bioi Rep
13:213-219, 1989
49. Elder JT, Fisher GJ, Lindquist PB, Bennett GL, Pittclkow MR,
Coffey RJ Jr, Ellingsworth L, Derynck R, Voorhees JJ:
Overexpression of transforming growth factor a in psoriatic
epidermis. SciCllce 243:811-814,1989
50. Gottlieb AB, Chang CK, Posnett DN, Fanelli B, Tam JP:
Detection of trans· forming growth factor-a in normal , malignant,
and hyperproliferative human keratinocytes.J Exp Med 167:670-675,
1988
51. Gottlieb A, Grossman R, Khandke L, Carter OM, Sehgal P, Fu
SM, Granelli-Piperno A, Rivas M, Barazani L, Krueger J: Studies of
the effect of cyclosporinr in psoriasis ill lIivo: combined effects
on activated T lymphocytes and epide.rn:ul regenerative
maturation.J [lIlIest DermatoI98:302 - 309, 1992
52. Cook PW, Pittelkow MR, Keeble WW, Graves-Deal R, Coffey RJ
Jr, Shipl. GD: Amphiregulin messenger RNA is elevated in psoriatic
epidennis and gastrointestinal carcinomas. Callcer Rcs
52:3224-3227, 1992
