Digestive Disea.ves and Sciences, Vol. 37, No. 6 (June 1992), pp. 81~26
`
`Tumor Necrosis Factor-a, Interleukin-lp,
`and Interleukin-6 Expression in
`Inflammatory Bowel Disease
`
`CHRIS STEVENS, MD, GERD WALZ, MD, CHANDER SINGARAM, MD,
`MARK L. LIPMAN, MD, BERND ZANKER, MD, ALDO MUGGIA, MD,
`DONALD ANTONIOU, MD, MARK A. PEPPERCORN, MD,
`and TERRY B. STROM, MD
`
`The etiology of ulcerative colitis (UC) and Grahn's disease (CD) remains enigmatic.
`Infiltrating intestinal macrophages are capable of producing the proinflammatory cyto(cid:173)
`kines tumor necrosis factor-a (TNF-a), interleukin-1 f3 (IL-l {3), and interleukin-6 (JL-6).
`We investigated the presence of IL-6, TN F-a and IL-l f3 mRNA transcripts in inflamma(cid:173)
`tory bowel disease (IBD), normal, and other inflammatory intestinal specimens utilizing
`the polymerase chain reaction (PCR). TNF-a mRNA levels did not vary between
`inflammatory bowel disease and control specimens. IL-l f3 mRNA levels were highest in
`active UC and noninflammatory bowel disease inflammatory specimens while /L-6 mRNA
`levels were highest in active IBD specimens.lnfiltrating Tcells, macrophages, and B cells
`were identified as sources of IL-6 protein in inflammatory bowel disease specimens by
`immunofluorescent staining. IL-6 transcripts were elevated only in active inflammatory
`bowel disease specimens, suggesting that IL-6-mediated immune processes are ongoing
`in the inflammatory mucosal environment of CD and UC.
`
`KEY WORDS: Crohn's disease; ulcerative colitis; polymerase chain reaction; interleukin-6; interleukin-1 ; tumor
`necrosis factor.
`
`The mucosal inflammatory response in IBD is
`thought to be either an appropriate reaction to an as
`yet unidentified antigen(s) or is the result of an
`aberrant immune system. Whatever the instigating
`factors, the observed result seen within the bowel
`wall is a prominant cellular infiltrate composed of
`
`Manuscript received June 17, 1991 ; revised manuscript re(cid:173)
`ceived December 18, 1991; accepted December 27, 1991.
`From the Department of Medicine, Center for Inflammatory
`Bowel Disease, Beth Israel Hospital and Harvard Medical
`School, Boston, Massachusetts 02215.
`Supported by grants from the NIH, Crohn's and Colitis
`Foundation of America, Inc., Medical Research Council of
`Canada, and Deutsche Forschungsgemeinschaft.
`Address for reprint requests: Dr. Chris Stevens, Beth Israel
`Hospital, Departments of Gastroenterology and Clinical Immu(cid:173)
`nology, Dana Bldg., Room 501 , 330 Brookline Ave., Boston,
`Massachusetts 02215.
`
`neutrophils, plasma cells, mast cells, lymphocytes,
`and macrophages. The activation products of these
`inflammatory cells contribute to the epithelial intes(cid:173)
`tinal damage and ultimate clinical disease. Eleva(cid:173)
`tion in prostaglandins (1), leukotrienes (2), active
`oxygen species (3), and the finding of anti-colon
`antibodies (4) and cytotoxic lymphocytes in the
`involved intestine have been identified in IBD pa(cid:173)
`tients (5). Peripheral blood monocytes isolated from
`IBD patients manifest increased motility and phago(cid:173)
`cytosis, and elaborate lysosomal enzymes, indicat(cid:173)
`ing that circulating monocytes are activated in IBD
`patients (6). Interleukin-I (IL-1), interleukin-6 (IL-
`6), and tumor necrosis factor (TNF) are three
`immune modulators termed cytokines produced by
`
`818
`
`Digestive Diseases and Sciences, Vol. 37, No. 6 (June /992)
`0163·2116ml()6()0-0818$1)6.5010 C 1992 Plenum Publishing Corporation
`
`

`

`CYTOKINES IN INFLAMMATORY BOWEL DISEASE
`
`TABLE I. INTESTINAL SPECIMENS ANALYZED FOR EACH CYTOKINE IN STUDY GROUPS
`
`Medicationt
`s
`
`c
`
`M
`
`2
`
`4
`
`4
`3
`0
`
`1
`3
`I
`2
`
`4
`2
`2
`
`Specimen
`
`IL-6
`
`JL-1 {3
`
`TN F-a
`
`SIE*
`
`Crohn's active
`Ileum
`Duodenum
`Crohn's inactive
`Ileum
`Colon
`UC active
`
`UC inactive
`Non-IBD inflammatory
`Esophagitis
`Gastritis
`Duodenitis
`Appendicitis
`Normals total
`Colon
`Esophagus
`Antrum
`Duodenum
`
`5
`4
`I
`6
`4
`2
`12
`
`7
`9
`3
`3
`2
`I
`10
`5
`2
`2
`I
`
`4
`3
`I
`4
`4
`0
`
`7
`2
`9
`3
`3
`2
`I
`6
`5
`1
`0
`0
`
`4/1
`
`0/1
`
`212
`012
`3/9
`3/4
`on
`1/8
`013
`0/3
`012
`1/0
`218
`213
`
`5
`4
`I
`6
`4
`2
`12
`
`7
`9
`3
`3
`2
`I
`10
`5
`2
`2
`1
`
`*Surgery/endoscopy.
`tC, corticosteroids; S, sulfasalazine; M, mesalamine enemas.
`
`macrophages whose functions have been elucidated
`over the past decade.
`IL-l, as well as IL-6, can provide a second signal
`for T -cell activation (7). Expression of IL-l sup(cid:173)
`ports B-cell activation and differentiation, augments
`NK cell cytolytic activity, and is directly toxic to
`certain cell lines (8). IL-l promotes fibrosis by
`stimul~ting fibroblast mitogenesis and increasing
`the transcription of type I, type Ill, and type IV
`collagen (9). When compared to normal controls,
`bioassay and ELISA measurements reveal that in(cid:173)
`creased levels of IL-l protein are present in intes(cid:173)
`tinal supernatants of cultured intestinal tissue and
`intestinal mononuclear cells extracted from IBD
`patients (10, 11).
`Distinct alpha and beta TNF polypeptides have
`been characterized (12). TN Fa promotes T-cell and
`B-cell proliferation and differentiation (13, 14).
`TNF-a is in~xtricably linked to the toxic shock
`syndrome (15) and has been detected in multiple
`·
`sclerosis plaques (16).
`IL-6 is a pleiotropic cytokine produced by mono(cid:173)
`cytes, macrophages, and also by activated T cells,
`B cells, fibroblasts, endotheHal cells, chondrocytes,
`and mesangial cells (17). IL-6 promotes terminal
`differentiation of B cells, induction of immunoglob(cid:173)
`ulin secretion (18), and production of acute-phase
`proteins by hepatocytes (19). IL-6 provides costim(cid:173)
`ulatory T-cell activation signals, which in turn sup(cid:173)
`port IL-2 mRNA transcription. IL-6 also promotes
`
`Digestive Diseases and Sciences, Vol. 37, No.6 (June 1992)
`
`the cytolytic capacity of cytotoxic T cells (20) and
`NK cells (21). IL-6 expression has been implicated
`in the pathogenesis of mesangial proliferative glo·
`merulonephritis (22), multiple myeloma (23), rheu(cid:173)
`matoid arthritis (24), and cardiac myxoma (25).
`Expression of TNF-cx, IL-1 (3, and IL-6 tran(cid:173)
`scripts have not been reported in intestinal speci(cid:173)
`mens obtained from IBD patients. Therefore we
`chose to investigate these <:;ytokine transcripts most
`likely to be produced by macrophages in the intes(cid:173)
`tinal mucosa of IBD patients.
`
`MATERIALS AND METHODS
`
`Experimental Design. Patients were selected for study
`from three categories: (1) IBD; (2) other gastrointestinal
`inflammatory diseases-esophagitis (N = 3), gastritis (N
`= 3), duodenitis (N = 2), appendicitis (N = 1); and (3)
`subjects lacking inflammatory or neoplastic diseases of
`the bowel (Table 1). In each case of CD. and UC the
`diagnosis was confirmed by clinical, radiographic, endo(cid:173)
`scopic, and histologic criteria (26). Non-IBD inflamma(cid:173)
`tory diagnoses were based upon endoscopic findings and
`histology revealing active inflammation. The study was
`approved by the committee on clinical investigation of
`Beth Israel Hospital. Informed consent was obtained
`from patients undergoing endoscopy. Surgkal resections
`and endoscopic biopsy samples were studied. Three
`endoscopic biopsies each were taken from normal area~
`and areas of gross inflammation in patients with IBD or
`other inflammatory diseases. Patients undergoing
`colonoscopy for cancer surveillance or esophagogas(cid:173)
`troduodenoscopy for evaluation of abdominal pain were
`
`819
`
`

`

`selected as normal controls. Similarly three biopsies were
`obtained from a grossly normi}l site in these patients.
`Specimens of bowel resected for neoplasia were< taken >
`8 em from any histologically abnormal lesions. O!)ly those
`specimens that manifested absolutely normal histology by
`hematoxylin and eosin staining without cryptitis, granu(cid:173)
`lomas, or increased inflammatory cell populations were
`included in the normal group. Giemsa stains were per(cid:173)
`formed on all normal tissues to specifically investigate the
`possibility of mast cell degranulation, which was not seen
`in any normal specimen. IBD and non-IBD inflammatory
`samples were interpreted as either moderately or severely
`active by a gastrointestinal pathologist.
`Two of the biopsies from each site were immediately
`frozen in liquid nitrogen and reserved for subsequent
`RNA extraction. The third biopsy was placed in Zambo(cid:173)
`ni's fixative and processed into frozen sections in prepa(cid:173)
`ration for immunofluorescent staining.
`Materials. Polyclonal rabbit anti-human IL-6 (Gen(cid:173)
`zyme, Boston, Massachusetts) and monoclonal mouse
`anti-human anti-CD3 , anti-CD22 (Becton Dickinson,
`Mountainview, California), and M0-2 (Coulter Immunol(cid:173)
`ogy, Hialeah, Florida) were used as the primary antibod(cid:173)
`ies in the immunofluorescent technique. Fluorescein iso(cid:173)
`thiocyanate (FITC) -conjugated goat anti-rabbit and
`rhodamine isothiocyanate (RITC) -conjugated goat anti(cid:173)
`mouse secondary antibodies were purchased from Cal(cid:173)
`biochem (La Jolla, California). Full-length IL-l~ and
`TNF-!X cDNAs were a generous gift from Charles
`Dinarello (Tufts University, Boston, Massachusetts) and
`IL-6 eDNA was kindly provided by Steve Clark (Genetics
`Institute, Cambridge, Mass~chusetts). Culture reagents
`included p~ytohemagglutinin (PHA, Wellcome Diagnos(cid:173)
`ti~s, Dartford, UK), phorbol myristate acetate (PMA,
`Sigma, St. Louis, Missouri), cyclohexamide (Sigma),
`RPMI 1640 media (MA Bioproducts, Walkerville, Mary(cid:173)
`land), normal human serum (Biobee, Boston, Massachu(cid:173)
`setts), HEPES buffer (Sigma), and Ficoll Hypaque (Phar(cid:173)
`macia Co., Piscataway, New Jersey). Guanidinium
`thiocyanate, sarcosyl, and cesium chloride were pur(cid:173)
`chased from International Biotechnologies (New Haven,
`Connecticut). Sodium citrate, sodium acetate, and ethyl(cid:173)
`enediamine tetraacetic acid (EDT A) were supplied by
`Fisher Scientific (Fair Lawn, New Jersey) while agarose,
`3-(4-morpholino)propane sulfonic acid (MOPS) and form(cid:173)
`amide were furnished by Sigma. The reducing agent
`2-mercaptoethanol was provided by Bio-Rad (Richmond,
`California) and Hybond N + nylon filter by Amersham
`(Arlington Heights, lllinois). Microscopic slides (Fisher
`Sc~entific) , gelatin (Biorad), normal goat serum (Gibco,
`Grand Island, New York), picric acid (Aldrich), and
`sucrose (M~linckrodt Inc. Paris, Kentucky) were em(cid:173)
`ployed in the immunohistochemical studies. The reverse
`transcriptase from Maloney murine leukemia virus (Be(cid:173)
`Uiesda Research Laboratories, Bethesda, Maryland) and
`the DNA polymerase of Thermus aquaticus (Perkin
`Elmer Cetus, Norwalk, Connecticut) were used for
`mRNA amplification.
`Tissue Sections. Samples were placed in Zamboni's
`fixative (2% formaldehyde, 15% picric acid in 0.1 M
`phosphate buffer) for 14-lg hr followed by soaking in
`phosphate-buffered saline (PBS), pH 7.4, at go C for 24
`820
`
`STEVENS ET AL
`
`hr. After an additional 24 hr at go C in PBS containing
`15% sucrose, the tissues were serially cryosectioned at 16
`1-Lm and placed on gelatin-coated slides.
`Immunofluorescent Staining. Two color indirect immu(cid:173)
`nofluorescence with the FITC-tagged IL-6 polyclonal
`antibody and RITC-Iabeled anti-CD3, anti-CD14, and
`anti-CD22 were utilized to demonstrate the cellular ori(cid:173)
`gins of IL-6 production within the bowel specimens.
`Non-antigen-specific antibody binding sites were blocked
`with 5% normal goat serum. Bowel sections were then
`incubated with the antibodies for 20 hr at 4o C. After
`washing the specimens with PBS, the secondary anti(cid:173)
`body, FITC-conjugated affinity-purified goat anti-rabbit
`IgG or RITC-conjugated goat anti-mouse IgG was applied
`for J hr at room temperature. Sections that were evalu(cid:173)
`ated for the cellular origins of IL-6 received an additional
`l-hr incubation at room temperature with the RITe(cid:173)
`conjugated antibody. Microscopic observation and pho(cid:173)
`tography were performed under fluorescent epiillumina(cid:173)
`tion on an Olympus BH-2 microscope.
`RNA Extraction. RNA was extracted from gut samples
`by the guanidinium isothiocyanate cesium chloride
`method (27). One to three milliliters of guanidinium
`thiocyanate (GTC) solution (4 M GTC, 25 mM sodium
`citrate, pH 7, 0.5% sarcosyl, and 2% 2-mercaptoethanol)
`was added to frozen ( - 80° C) tissue. DNA was sheared
`with a polytron (Brinkmann, Westbury, New York). The
`samples were layered upon a 5.7 M cesium chloride
`gradient and ultracentrafuged for }g hr at 35,000 rpm. The
`RNA pellet was recovered, ethanol precipitated, and
`quantitated by absorbance at 260 nm.
`Fifteen to 20 micrograms of total RNA were routinely
`extracted from two endoscopically obtained biopsies, and
`the amounts were comparable in normal and inflamma(cid:173)
`tory tissue. Two micrograms of total RNA from all
`samples were electropheresed on 1% agarose gels and
`stained with ethidium bromide to identify 28S and Jgs
`riposomal bands. This ensured the integrity of the ex(cid:173)
`tracted RNA prior to proceeding with subsequent steps in
`gene amplification.
`Northern RNA Transfer Protocol. RNA samples were
`dissolved in 50% formamide, 2.2 M formaldehyde, 40 mM
`MOPS, 10 mM sodium acetate, I mM EDTA, and dena(cid:173)
`tured by heating for 5 min at 60° C and applied to a I%
`agarose gel prepared using standard methods. Twenty
`micrograms of RNA were electropheresed for 4 hr at 60 V
`(go rnA) and then transferred to Hybond N + nylon filters
`according to the recommendations of Amersham. After
`air drying, the filters were hybridized overnight at 42° C
`in a solution containing 50% formamide, 5X SSC, 50 mM
`sodium phosphate, pH 6.5, 0.2% SDS, J x Denhardt's
`solution, 10% dextran, 100 f.Lg/ml salmon sperm DNA,
`and eDNA probe radiolabeled by the random primer
`method. After hybridization, the filter was washed at
`room temperature with 2x SSC and 0.1% SDS for 30 min
`followed by another 30 min at 55° c in a 0.2x sse, 0.1%
`SDS solution. After air drying, the filters were exposed to
`x-ray film (Kodak, Rochester, New York) for 24- 72 hr.
`Synthesis of eDNA and Polymerase Chain Reaction
`(PCR) Amplification. Oligonucleotide sense TTAAGCT(cid:173)
`TGCTATGAACTCCTTCTCCACAAGC, TTAAGCTT(cid:173)
`TGGCTGAACCGCCGGGCAATGCC,TTAAGCTTGC-
`
`Digestive Diseases and Sciences, Vol. 37, No. 6 (June /992)
`
`

`

`CYTOKINES IN INFLAMMATORY BOWEL DISEASE
`
`CATGGACAAGCTGAGGAAGATG and antisense
`CAGGATCCCATGCTACATTTGCCGAAGAGCCCT,
`CAGGATCCTCACAGGGCAATGATCCCAAAGTA,
`CAGGATCCTCTTT AGGAAGACAAA TTGCAT prim(cid:173)
`ers complementary to IL-6, TNF-a , and IL-113 DNA
`sequences, respectively, were prepared using a DNA
`synthesizer (Applied Biosystems, Foster City, Califor(cid:173)
`nia). A single-strand eDNA copy was made from 1 f.J-g of
`total RNA using 5 mM of the appropriate antisense
`oligonucleotide primer for each eDNA synthesis.
`Amplification of the first strand of eDNA was per(cid:173)
`formed with the heat-stable DNA polymerase of Thermus
`aquaticus as recommended by the supplier (Perkin Elmer
`Cetus). Twenty, 25, 30, and 40 cycles of PCR amplifica(cid:173)
`tion were performed using a DNA thermal cycler (Perkin
`ELmer Cetus). This established the linear part of the
`ampLification curve for each cytokine (28). Each cycle
`consisted of 1 min at 94° C to denature double-stranded
`DNA, 45 sec at 60° C for the primers to anneal to their
`complementary sequences, and I min at 72o C for exten(cid:173)
`sion of the DNA strands. The resulting PCR products
`were size fractionated on 1.5% agarose gels, bloued onto
`nylon membranes and hybridized with random primer
`radiolabeled full-length cytokine cDNAs. Total RNA
`extracted from PHA (5 f.J-g/ml) and PMA (5 ng/ml) stimu(cid:173)
`lated peripheral blood mononuclear cells (PBMC) cul(cid:173)
`tured for 12 hr and pulsed with cyclobexamide (20 f.J-g/ml)
`for an additional4 hr served as the positive control for the
`detection of IL-6, TNF-a, and IL 113 transcripts in each
`reaction. The same stock of positive control RNA was
`used for each set of amplification reactions and produced
`a strong and consistent hybridization signal with TNF-a,
`IL-113, and IL-6 probes. Reaction mixtures containing
`oligonucleotides, appropriate buffers, and enzymes with(cid:173)
`out RNA substrate served as the negative control for each
`reverse transcriptase and gene amplification reaction.
`Data Analysis. Semiquantitation of the PCR products
`was assessed by comparing the densitometric ratio of
`intestinal and control sample autoradiograms. Statistical
`analysis of the densitometry data was calculated by the
`analysis of variance (ANOV A).
`
`RESULTS
`
`Detection of Cytokine mRNA in Intestinal Speci(cid:173)
`mens. In an effort to detect cytokine mRNAs in
`intestinal samples, Northern blot analysis was ini(cid:173)
`tiaJJy employed . Cytokine transcripts for total RNA
`were not identified via this method in normal or
`inflammatory intestinal s p eci me ns (data not
`shown). Given the small amount of RNA extracted
`from endoscopic biopsy specimens, PCR gene am(cid:173)
`plification methods were utilized to detect cytokine
`transcripts and to determine whether qualitative
`differences in the tissue levels of cytokine tran(cid:173)
`scripts were evident among the various intestinal
`specimens analyzed. Specific cytokine PCR prod(cid:173)
`ucts were identified by hybridization with radiola(cid:173)
`beled full-length cytokine cDNAs {Figure 1).
`
`Digestive Oisease.r and Sciences. Vol. 37. No. 6 (June 1992)
`
`A
`POLYMERASE~ REACTION
`
`-z
`
`:
`:
`z
`~
`z
`z
`...
`...
`...
`0
`0
`0
`0
`0
`0
`u
`u
`u
`u ~ g
`
`:>
`
`z
`0
`
`..
`
`0
`u
`j
`
`:
`~
`..
`..
`..
`..
`z
`z
`• !
`:1
`,..
`,..
`0
`0
`:>
`0: = = u
`...
`0
`0
`,..
`,..
`0:
`z
`g g
`~
`c
`
`:1
`:>
`
`0:
`u
`
`15
`
`TN Fa ..... IIIJ1-- ---- 407bp
`
`~
`
`IL-1 b
`
`622bp
`
`Fig JA. PCR products of intestinal samples size fractionated on
`1.5% agarose gels and hybridized with radiolabeled probes for
`(A) TNF-a, IL- l~ and (B) IL-6. TNF-a was amplified 30 cycles
`(PCR products were not seen at 25 cycles; data not shown),
`while lL-1(3 and lL-6 were amplified 40 cycles each. NL, UC,
`and CR abbreviations stand for normal, ulcerative colitis, and
`Crohn's disease specimens, respectively.
`
`Measurement of the relative densities of the hy(cid:173)
`bridized PCR products provided a means for esti(cid:173)
`mating the relative abundance of intestinal cytokine
`mRNAs (Figure 2). TNFa was detected in all intes(cid:173)
`tinal samples studied after 30 cycles of amplifica(cid:173)
`tion, and the magnitude of expression determined
`by the percent densitometry o f control did not vary
`to a statistically significant degree among inflamma(cid:173)
`tory and normal groups. Expression of lL-1~ was
`pronounced in active UC with relative densities of
`the PCR product in active UC s pecimens greater
`than those detected in active CD, inactive IBD, and
`normal specimens (P < 0.05). However, the differ(cid:173)
`ence between active UC and non-IBD inflammatory
`samples was not statistically significant. IL-6
`mRNA levels were much higher in IBD specimens
`than inactive IBD, non-IBD inflammatory, and nor-
`821
`
`

`

`STEVENS ET AL
`
`.J
`
`z
`
`-
`
`II:
`
`(,)
`
`(,)
`
`;;,
`
`;;,
`
`POLYMERASE CHAIN REACTION
`c
`...
`:I
`0
`.J
`z
`:I
`z
`;;,
`z
`...
`(,)
`:I
`;;, z z z z
`c
`• • Q
`-
`-
`c
`:I ~ 0
`0
`0 0
`...
`- - ;;,
`...
`;;,
`... 0
`...
`(,)
`(,)
`.J
`.J
`~ 0
`.J .J
`Q - .J ~ ...
`(,)
`0
`0
`0 0
`-
`c
`(,) z
`II:
`(,)
`(,)
`I
`~ ~
`~ z 2
`15 z (,) g (,) ~ ~
`15
`c
`
`.J
`z
`
`:I
`;;,
`
`.J
`
`c
`(,)
`
`:I
`;;,
`
`:I
`;;,
`
`~
`
`(,)
`
`...
`...
`
`A.
`
`A.
`
`ll- 6
`
`633bp
`
`Fig lB. A representative blot of the 633bp PCR product for IL-6 from intestinal tissues is
`shown after electropheresis on a 1.5% agarose gel and hybridization with a random labeled
`IL-6 eDNA. The positive control (PIP/C) is a sample of peripheral blood mononuclear cells
`stimulated with PHA, PMA, and cyclohexamide. NL, UC, and CR abbreviations stand for
`normal, ulcerative colitis, and Crohn's disease specimens, respectively.
`
`mal groups (P < 0.05). The application of cortico(cid:173)
`steroid, sulfasalazine, or 5-aminosalacylic acid
`treatment to the data did not correlate with the
`amount of IL-6 mRNA detected in IBD specimens.
`Detection of IL-6 Protein in IBD Specimens. To
`determine whether the IL-6 transcripts, detectable
`
`only by very sensitive PCR techniques, were
`translated into protein products, indirect immun·
`oftuorescent staining of IL-6 was undertaken using
`polyclonal anti-IL-6 antibodies. Representative
`sections of active and healed ulcerative colitis
`samples (Figure 3) clearly demonstrate cytoplas-
`
`DENSITOMETRY OF IL-6, IL-1b, AND TNFa
`PCR PRODUCTS
`
`140
`
`100
`
`;?
`~ 120
`~
`U5
`z
`UJ
`0
`UJ >
`F
`<t:
`...J
`UJ a:
`
`80
`
`60
`
`40
`
`• UCACTIVE
`• CD ACTIVE
`
`0 UCINACTIVE
`1?:1 CD INACTIVE
`0 NON·IBD ACT.
`l2l NOR\1AL
`
`20
`
`0
`
`IL-6
`TN Fa
`IL-1b
`Fig 2. Densitometry scanning of lL- 113, TNF-o:, and IL-6 PCR products expressed as a percent of
`the positive control. The same control was used for all experiments to obtain these comparative
`results. All specimens in Table I are represented in this figure. The relative density oflL-6 from UC
`and CD active samples was higher (*P < 0.05) than all the other groups. IL-1~ in UC active
`specimens was greater than CD active, CD inactive, and normal groups (*P < 0.05), but did not
`reach statistical significance when ~ompared to UC inactive and non-IBD active inflammatory
`specimens. TNF-a was present io all samples and the relative density did not vary significantly
`among the groups studied. Bars indicate + or -
`the standard error.
`
`822
`
`Digestive Diseases and Sciences, Vol. 37, No. 6 (June 1992)
`
`

`

`CYTOKINES IN INFLAMMATORY BOWEL DISEASE
`
`A
`
`IMMUNOFLUORESCENT STAINING
`c
`
`E
`
`IL - 6
`
`SURFACE
`MARKERS
`
`B
`
`D
`
`F
`
`H
`
`uc
`
`uc
`
`uc
`
`uc
`HEALED
`Fig 3. Representative sections of active and inactive UC colon specimens stained using indirect immunofluorescence for
`IL-6 (A, C, E , G) and macrophage CD14 (B, H), T-cell CD3 (D), and B-cell CD22 (F) surface markers are depicted. The
`cytoplasm of inflammatory cells stained brightly for IL-6 (green) in the active (A, C, E) but not the healed (G) UC sample.
`Simultaneous staining (red) identified macrophages (B), T cells (D), and B cells as sources of IL-6. The healed UC section
`(H) contained macrophages not producing lL-6. Solid straight arrows identify colocalized cells staining positively for both
`IL-6 and a surface marker. Open arrows localize those cells that are positive for surface markers only. The curved arrow
`identifies an artifact. The bar equals 50 !Lm.
`
`mic granular fluorescence in inflammatory cells of
`active colitis. Anti-IL-6 fluorescence was evident
`only in the active, but not the healed, UC colon
`sections. IL-6 protein was not identified within
`epithelial, fibroblast, or vascular endothelial cells.
`Colocalization with anti-CD3, anti-CD14, and an(cid:173)
`ti-CD22 monoclonal antibodies demonstrated that
`T cells, macrophages, and B cells, respectively,
`each synthesized IL-6 within the inflamed bowel
`wall. Similar results were found in the active
`Crohn's disease specimens. IL-6 was not identi(cid:173)
`fied in non-IBD specimens. These data verify that
`the IL-6 mRNA, detected indirectly by the PCR, is
`translated into protein by a varied group of white
`blood cells infiltrating the bowel wa)] of IBD
`patients.
`
`Digestive Diseases and Sciences, Vol. 37, No. 6 (June 1992)
`
`DISCUSSION
`
`We have used PCR gene amplification tech(cid:173)
`niques to identify cytokine transcripts, ie, IL-6,
`TNF-a, and IL-l~ mRNAs, in surgically resected
`and endoscopically obtained intestinal tissues.
`This powerful gene amplification technique was
`utilized because application of conventional
`Northern blot detection methods to intestinal bi(cid:173)
`opsy samples failed to reveal cytokine transcripts.
`The range of samples studied included esophageal,
`gastric, duodenal, ileal, appendiceal, colonic, and
`rectal specimens. Our study revealed that TNF-a
`mRNA was detectable in each specimen examined
`and is therefore probably constitutively expressed
`in the esophagus, stomach, ileum, and colon;
`however, tissue levels of TNF-a transcripts were
`823
`
`

`

`not increased in IBD specimens. Others have not
`detected TNF-cx by ethidium bromide staining of
`PCR products from normal or active IBD speci(cid:173)
`mens (29). We also could not discern TNF-cx in
`many samples with ethidium bromide analysis
`alone, yet when hybridized with a radiolabeled
`TNF-a eDNA, which markedly enhances sensitiv(cid:173)
`ity, specific products became readily visible in all
`samples.
`IL-l~ mRNA was present in most inflammatory
`samples, but absent in normal specimens, suggest(cid:173)
`ing that expression ifiL-1~ mRNA if translated and
`secreted supports the inflammatory processes of
`esophagitis, gastritis, appendicitis, and IBD. Lower
`levels of IL-l~ mRNA were identified in active CD
`specimens as compared to active UC samples. This
`observation may reflect the difference between
`small intestinal and colonic inflammatory cell pop(cid:173)
`ulations since all CD active samples were from the
`small bowel. Alternatively, enterocyte production
`of IL-l~, which has been demonstrated in experi(cid:173)
`mental colitis (30), may account for the difference
`since the mucosa of UC is diffusely inflamed in
`comparison to the transmural involvement in CD.
`IL-l~ mRNA levels did not differ statistically be(cid:173)
`tween active and inactive ileal CD. This finding
`correlates with supernatant IL-l~ ELISA measure(cid:173)
`ments of sonicated ileal samples (31). The discovery
`of IL-l~ mRNA in active UC mucosa is consistent
`with the data of others, who have detected elevated
`levels of IL-l~ protein in supernatants of UC tissue
`and intestinal mononuclear ce1ls extracted from
`active UC specimens when compared to normal
`controls (10, 11).
`The profile of IL-6 mRNA tissue accumulation
`was of particular interest. Within the range of
`biopsy materials examined, IL-6 mRNA levels are
`increased over normals and the non-IBD specimens
`studied. IL-6 mRNA positive IBD specimens were
`also positive for IL-l~ and TNF-cx mRNA. There(cid:173)
`fore, expression of IL-6 may provide an additional
`inflammatory mediator in IBD that, in conjunction
`with IL-1~, TNF-a, and other inflammatory mole(cid:173)
`cules, leads to the inflammatory response charac(cid:173)
`teristic of active IBD, or IL-6 may play a protective
`role by inducing hepatic acute phase proteins (19).
`As IL-6 was detected in IBD samples only by
`PCR but not Northern blot, given the limited
`amount of RNA obtainable from biopsy tissues, we
`were interested to determine whether IL-6 mRNA
`was translated into measurable protein, and if so
`which of the many cell types capable of synthesiz-
`
`824
`
`STEVENS ET AL
`
`ing IL-6 were responsible for its production in IBD
`specimens. Immunofluorescent techniques identi(cid:173)
`fied IL-6-positive cytoplasmic granules only in in(cid:173)
`flammatory ceHs of IBD sections. IL-6 was not
`detected in inactive IBD or non-IBD inflammatory
`samples. Colocalization studies utilizing anti-lL-6
`and cell-specific surface phenotypic markers iden(cid:173)
`tified macrophages, B cells, and T-cells as potential
`sources of IL-6. From these data it is unknown
`whether the cytoplasmic IL-6 detected is secreted
`into the intestinal microenvironment. Other inflam(cid:173)
`matory cells, including mast cells, plasma cells, and
`NK cells, that were not studied may also be pro(cid:173)
`ducing IL-6.
`TNF-a and IL-1~ can both induce expression of
`IL-6 in vitro, and both are present in non-IBD
`inflammatory samples; yet IL-6 is undetectable.
`Since IL-l can induce expression of IL-6, the local
`concentration of TNF-a and IL-1 present in non(cid:173)
`IBD inflammatory tissues may be insufficient to
`induce the expression of IL-6. Alternatively, other
`stimulants may induce IL-6 in IBD. For example,
`viruses can induce IL-6 transcription via a cis(cid:173)
`acting mechanism (32). Perhaps an as yet unidenti(cid:173)
`fied etiologic agent causing IBD stimulates the
`transcription or inactivates a newly identified IL-6
`repressor mechanism (33). The proinflammatory
`actions ofTNF-cx, IL-1~ , and IL-6 are all likely to
`be involved in the mucosal inflammation of this
`disorder. In particular, the activities of IL-6 corre(cid:173)
`late well with the immunological and clinical mani(cid:173)
`festations of IBD and other "autoimmune" dis(cid:173)
`eases. Polyclonal B-cell act ivation (34),
`immunoglobulin production (35), and T-een prolif(cid:173)
`eration and differentiation (36) have been described
`iniBD.
`Our data identify cytokine transcripts and cyto(cid:173)
`plasmic IL-6 protein in active IBD tissue. In partic(cid:173)
`ular the presence of JL-6, found only in IBD spec(cid:173)
`imens as well as IL-l~ and TNF-cx, which are also
`expressed in IBD, may promote local inflammatory
`effects leading to IBD disease expression. Hence
`effective treatments for IBD should block IL-6
`production. Indeed corticosteroids can block IL-l
`(37) and IL-6 synthesis at the transcriptional level
`and prevent the production of many other cytokines
`(38-41). Although it is unlikely that the inhibition of
`IL-6 alone explains the effectiveness of corticoster(cid:173)
`oids in the treatment of IBD, the inhibitory effect of
`corticosteroids on expression of .IL-6 and other
`cytokines does lend support to the involvement of
`cytokines in the pathogenesis of IBD.
`
`Digestive Diseases and Sciences, Vol. 37, No. 6 (June 1992)
`
`

`

`CYTOKINES IN INFLAMMATORY BOWEL DISEASE
`
`In summary, the cytokine transcripts encoding
`IL-6, TNF-a, and IL-113 were detected by PCR in
`inflammatory intestinal tissues. Of the inflammatory
`samples studied-esophagitis, gastritis, duodenitis,
`appendicitis, UC, and CD-IL-6 was predomi(cid:173)
`nantly detected in UC and CD specimens. Immun(cid:173)
`ofluorescent techniques demonstrated IL-6 protein
`in T cells, B cells, and macrophages. These data
`suggest that IL-6 may be an intestinal inflammatory
`mediator of both UC and CD.
`
`ACKNOWLEDGMENTS
`We would like to thank Sybil Goulkin, the staff of the
`endoscopy suite, and the pathology department of the
`Beth Israel Hospital for their assistance in obtaining and
`processing intestinal specimens.
`
`REFERENCES
`
`I. Donowitz M: Arachadonic acid metabolites and their role in
`inflammatory bowel disease. Gastroenterology 88:580, 1985
`2. Sharon P, Stenson WF: Enhanced synthesis of leukotriene
`B4 by colonic mucosa in inflammatory bowel disease. Gas(cid:173)
`troenterology 86:453, 1984
`3. Kitahora T, Suzuki K, Asakura H, Yoshida T, Suematsu,
`Watanabe M, Aiso S, Tsuchiya M: Active oxygen species
`generated by monocytes and polymorphonuclear cells in
`Crohn's disease. Dig Dis Sci 33:951-955, 1988
`4. Broberger 0, Perlmann P: Autoantibodies in human ulcer(cid:173)
`ative colitis. J Exp Med 110:657, 1959
`5. Kirsner JB, Shorter RG: Inflammatory Bowel Disease. Phil(cid:173)
`. adelphia, Lea & Febiger, 1988
`6. Whorwell PJ, Bennett P, Tanner AR, Wright R: Monocyte
`function in Crohn's disease and ulcerative colitis. Digestion
`22:271-275, 1981
`7. Kurt-Jones EA, Kiely JM, Unanue ER: Conditions required
`for expression of membrane IL-l on B-<:ells. 1 Immunol
`135:1548-1550, 1985
`8. Dinarello CA: Interleukin-1 and its biologically related cy(cid:173)
`tokines. Adv Immunol44:153-205, 1989
`9. Krane S, Dayer JM, Simon CS, Byrne S: Mononuclear
`cell-conditioned medium containing mononuclear cell factor
`(MCF), homologous with interleukin J, stimulating collagen
`and fibronectin synthesis by adherent rheumatoid synovial
`cells: effects of prostaglandin E2 and indomethacin. Collagen
`Rei Res 5:99-117, 1985
`10. Mahida YR, Wu KC, Jewell DP: Enhanced production of
`interleukin !-beta by mononuclear cells isolated from mu(cid:173)
`cosa with active ulcerative colitis or Crohn's disease. Gut
`30:835-838, 1989
`I I. Ligumsky M, Simon PL, Karmeli F. Rachmilewitz D: Role
`of interleukin I in inflammatory bowel disease-enhanced
`production during active disease. Gut 31:686-689, 1990
`12. Pennica D, Nedwin GE, Haytlick JS, Seeburg PH, Derynck
`R, Palladino MA, Kohr WJ, Aggarwal BB, Goedde! DV:
`Human tumor necrosis factor: Precursor structure, expres(cid:173)
`sion and homology to lymphotoxin. Nature 312:724-729,
`1984
`
`Digestive Diseases and Sciences. Vol. 37. No. 6 (June /992)
`
`13. Scheurich P, Thoma B, Ucer U, Pfizenmaier K: Immuno(cid:173)
`regulatory activity of recombinant human tumor necrosis
`factor (TNF)a: Induction o