ARTHRITIS & RHEUMATISMVol. 48, No. 10, October 2003, pp 2796–2806DOI 10.1002/art.11253© 2003, American College of Rheumatology

Synovial Fluid Neutrophils Transcribe and Express Class IIMajor Histocompatibility Complex Molecules

in Rheumatoid Arthritis

Andrew Cross,1 Roger C. Bucknall,1 Marco A. Cassatella,2 Steven W. Edwards,1

and Robert J. Moots1

Objective. To investigate a potential interactionbetween neutrophils and T cells in rheumatoid arthritis(RA), by defining the optimal conditions for induction ofclass II major histocompatibility complex (MHC) ex-pression on peripheral blood neutrophils in vitro andinvestigating the capacity for neutrophils to expressclass II MHC molecules in RA.

Methods. Surface expression of class II MHC andcostimulatory molecules by peripheral blood and syno-vial fluid (SF) neutrophils obtained from healthy con-trols and patients with RA was measured by flowcytometry and fluorescence microscopy. Intracellularclass II MHC protein and messenger RNA (mRNA)were detected by Western blotting and Northern blot-ting, respectively.

Results. Freshly isolated peripheral blood neutro-phils from controls did not express surface class IIMHC; expression was induced by culture with appro-priate cytokines. Freshly isolated peripheral blood neu-trophils from patients with RA expressed mRNA, butthere was no surface expression of class II MHC.Freshly isolated SF neutrophils from patients with RAcontained high levels of class II MHC mRNA, did notexpress surface class II MHC, but did have largeintracellular amounts of this protein as detected byWestern blotting. After culture for 20 hours in vitro, SF

neutrophils from RA patients expressed large amountsof surface class II MHC but very low levels of costimu-latory molecules CD80 and CD86. Fluorescence micros-copy localized surface class II MHC to discrete areas onthe neutrophil. Class II MHC–expressing neutrophilsstimulated T cell proliferation.

Conclusion. Peripheral blood neutrophils frompatients with RA but not healthy controls express classII MHC mRNA. SF neutrophils in RA synthesize andexpress large amounts of class II MHC but not costimu-latory molecules. This might underlie a novel interac-tion with T cells that is important in terms of diseasepathology.

The pathologic processes underlying rheumatoidarthritis (RA) remain unclear. Although T lymphocytesare believed to play an important role in this disease(1–6), anti–T cell therapies have generally failed to fulfilltheir potential in the clinical setting (7), suggesting thatother factors are important. Such factors might mediatedisease directly or through modulation of T cell func-tion.

In synovial fluid (SF) of the rheumatoid joint, themost abundant cells by far are neutrophils, which com-prise as much as 80% of all infiltrating cells (8,9) and arepresent in high numbers at sites of bone erosion. In thesetting of RA, neutrophils have been subjected to lessintensive scrutiny than have T cells, yet neutrophils haveconsiderable potential to contribute to damage in in-flamed joints by virtue of their ability to secrete toxicproteases and reactive oxygen intermediates (10,11), andthey are important in animal models of inflammatoryarthritis (12). In the past, neutrophils were considered tobe terminally differentiated at sites of inflammation andto survive there for only short periods of time.

We and other investigators, however, have clearly

Supported by the Arthritis Research Campaign, UK, and theUniversity of Liverpool.

1Andrew Cross, BSc, PhD, Roger C. Bucknall, MBBS,MRCP, FRCP, Steven W. Edwards, BSc, PhD, Robert J. Moots,MBBS, PhD, FRCP: University of Liverpool, Liverpool, UK; 2MarcoA. Cassatella, MD: University of Verona, Verona, Italy.

Address correspondence and reprint requests to Steven W.Edwards, BSc, PhD, Biosciences Building, School of Biological Sci-ences, University of Liverpool, Liverpool L69 7ZB, UK. E-mail:S.W.Edwards@liv.ac.uk.

Submitted for publication April 24, 2002; accepted in revisedform June 13, 2003.

2796

shown not only that neutrophils persist for much longerperiods of time upon inflammatory activation (13–15),but also that they have the capacity to up-regulate andexpress a variety of genes (16), including components ofthe NADPH oxidase, some cell surface receptors (e.g.,Fc� receptor type I [Fc�RI]), and a variety of cytokines(17–19). We have also reported that neutrophils isolatedfrom the SF of patients with RA express Fc�RI on theircell surface, indicating that activated de novo biosynthe-sis occurs in vivo during disease (14,20). It is thus clearthat neutrophils can respond to factors within their localenvironment to change their molecular properties andhence acquire the capacity to perform new cellularfunctions.

Recently, reports have shown that, when culturedwith appropriate cytokines in vitro (such as granulocyte–macrophage colony-stimulating factor [GM-CSF] andinterferon-� [IFN�]), mature neutrophils can be inducedto acquire characteristics normally found only on pro-fessional antigen-presenting cells, such as expression ofclass II major histocompatibility complex (MHC) (21).Furthermore, neutrophil precursors (band cells) can becultured in vitro with GM-CSF, tumor necrosis factor �,and interleukin-4 (IL-4) to acquire molecular and mor-phologic features of dendritic cells, including the expres-sion of class II MHC and costimulatory molecules suchas CD80 (B7-1) (22,23). These observations are partic-ularly important, because these cytokines are found insimilar concentrations in the inflamed joints of patientswith RA (24,25). Indeed, if neutrophils within diseasedjoints in RA were able to express such molecules, theywould have the potential to present antigen to T cells.Such presentation of antigen by nonprofessionalantigen-presenting cells might modulate T cell functionin a variety of ways, ranging from exposing crypticepitopes that would not normally have been exposedthrough breakdown of proteins in endosomes of den-dritic cells, to antigen-specific inactivation of T cells ifthere was signaling through a T cell receptor in theabsence of costimulatory molecules. Any of these mech-anisms would be of potential importance in the media-tion of chronic inflammation in RA.

In this study we investigated the capacity forneutrophils to express class II MHC molecules, bydefining the conditions for maximal induction of class IIMHC expression on peripheral blood neutrophils invitro. We also determined whether neutrophils isolatedfrom the SF of patients with RA can synthesize andexpress class II MHC and costimulatory molecules. Thepotential capacity of neutrophils to modulate T cellfunction, whether by stimulation or suppression, might

be an important factor underlying the abnormal, pro-longed immune response that is so damaging in RA.

MATERIALS AND METHODS

Materials. Neutrophil isolation medium (NIM) wasobtained from Cardinal Associates (Santa Fe, NM), Ficoll-Paque was obtained from Amersham Pharmacia (Uppsala,Sweden), and RPMI 1640 medium was obtained from GibcoBRL (Paisley, UK). Fetal bovine serum and Quantum calibra-tion microbeads were obtained from Sigma (Poole, UK). Theantibodies used were as follows: L243 (anti–HLA–DR, anti-genic groove; American Type Culture Collection, Rockville,MD), DA6.147 (anti–HLA–DR, C-terminal of cytoplasmic tailof � subunit; European Collection of Cell Cultures [ECACC],Poole, UK), DA6.231 (anti–HLA–DR, external region of �subunit; ECACC), anti-CD15 (Dako, Ely, UK), anti-CD16(Becton Dickinson, Oxford, UK), and anti-CD80 and anti-CD86 (PharMingen, Cowley, UK). The cytokines used were asfollows: IFN� (Roche Diagnostics, Lewes, UK), GM-CSF(Glaxo, Greenford, UK), and IL-3 (R&D Systems, Abingdon,UK).

Cell isolation and culture. This study was approved bythe local institutional review board, South Sefton ResearchEthics Committee. Peripheral blood cells were obtained fromheparinized venous blood from healthy donors, and frompatients with RA (according to the 1987 revised criteria of theAmerican College of Rheumatology [formerly, the AmericanRheumatism Association]) (26) and other inflammatory jointarthropathies, and were separated into neutrophil and mono-nuclear cell fractions by using NIM (according to the manu-facturer’s instructions) (27). Contaminating erythrocytes wereremoved by hypotonic lysis. Neutrophils from the SF ofpatients with RA were isolated soon after aspiration by usingFicoll-Paque in a 1-step centrifugation method (28).

Using May–Grunwald–Giemsa staining and trypanblue exclusion, neutrophils were routinely examined for purityand viability, which were �97% and �98%, respectively,immediately after isolation. Purity was confirmed using mor-phologic analysis of cytospins, and CD15 and/or CD16 expres-sion. The Raji B lymphoma cell line was used as a positivecontrol for class II MHC expression. Purified neutrophils wereresuspended in RPMI 1640 supplemented with 10% fetal calfserum (FCS) at 5 � 106 cells/ml and cultured at 37oC in ahumidified CO2 chamber. Cytokines were added as describedbelow. Conditioned media from Staphylococcus enterotoxin B(SEB)–stimulated monocytes was used as a source of mixedcytokines (29,30). Unfractionated SF was also cultured at 37oCin a humidified CO2 chamber.

T cell proliferation assays. Neutrophils isolated fromthe blood and SF of patients with RA were purified asdescribed above. Blood monocytes were isolated from themononuclear cell layer obtained after NIM centrifugation, byadherence to plastic. CD4� T cells were isolated by positiveselection using MACS CD4 MicroBeads and a separationcolumn (Miltenyi Biotec, Bisley, UK). Neutrophils were incu-bated overnight in the presence of GM-CSF (50 units/ml) andIFN� (100 units/ml) at 37oC to induce class II MHC expres-sion. Neutrophils and monocytes were then incubated for 3hours at 37oC with SEB (250 ng/ml) and then washed to

INDUCIBLE EXPRESSION OF CLASS II MHC IN INFLAMMATORY NEUTROPHILS 2797

remove excess superantigen. Cells (105) were then plated into96-well plates, and then 105 CD4� T cells were added. Plateswere incubated for 24 or 48 hours at 37oC (as indicated in thefigure legends), and then tritiated thymidine (1 �Ci/well) wasadded, followed by incubation for an additional 3 hours. Cellswere then harvested onto UniFilter plates (Perkin Elmer,Cheshire, UK), MicroScint-O (Perkin Elmer) was added, andincorporated radioactivity was measured using a TopCountScintillation Counter (Perkin Elmer). In some experiments,L243 antibody (at 100 ng/ml) was added to the incubationmixtures in order to confirm the role of class II MHCmolecules in the T cell proliferation.

Flow cytometry. Freshly isolated and cultured cellswere centrifuged at 2,000 revolutions per minute (450g), andthe supernatant was discarded. Cells were resuspended at 1 �106 cells/200 �l in blocking buffer (phosphate buffered salinecontaining 2 mg/ml bovine serum albumin) and then stainedfor flow cytometry using a standard indirect immunostainingassay (31). After incubation for 30 minutes at 4oC with asaturating concentration of the appropriate primary antibody,cells were washed and incubated with saturating fluoresceinisothiocyanate (FITC)–labeled F(ab�)2 goat anti-mouse sec-ondary antibody. Stained cells were then examined on an EpicsAltra flow cytometer (Beckman Coulter, Krefeld, Germany).A total of 10,000 cells per sample was analyzed. Relativebinding was measured by comparing the mean fluorescenceintensity of positive cells and that of negative (isotype control)cells. Calibration microbeads were used to determine therelative number of class II MHC molecules on the neutrophilcell surface.

Western blot analysis. After incubations as describedabove, cell pellets were resuspended in boiling Laemmli sam-ple buffer (1 � 106 cells/20 �l), vortexed, and boiled for 5minutes (32). Samples were then stored at �80oC until re-quired and electrophoresed on a 12% resolving polyacryl-amide gel used with a 4.5% stacking gel. Proteins were thenelectrophoretically transferred to polyvinylidene difluoridemembrane. After incubation with primary and secondaryantibodies, bound antibody was detected using an ECL West-ern Blotting Detection Kit (Amersham). Densitometry fromcarefully exposed blots (to avoid film saturation) was per-formed with NIH Image 1.44 VDM software (National Insti-tutes of Health, Bethesda, MD). The presence of ponceauS–stained actin on membranes confirmed equivalence ofloading.

Northern blot analysis. Total RNA was extracted from3 � 107 neutrophils or 1 � 106 HLA–DR–positive Raji humanB lymphoblastoid cells (33), using TRIzol (34) or guanidiniumisothiocyanate/CsCl2 (35). Total RNA was separated by form-aldehyde gel electrophoresis before capillary transfer toZetaProbe GT (Bio-Rad, Hercules, CA) membranes. Probeswere labeled with 32P-dCTP using a random primed labelingkit (High Prime; Roche Diagnostics, Mannheim, Germany).Membranes were sequentially probed, and radioactivity wasdetected and quantified using a Molecular Imager GS363(Bio-Rad).

Confocal microscopy. Cells were immunostained forexpression of class II MHC molecules, as described above, andthen propidium iodide was added to 10 �g/ml for 10 minutes tostain chromatin. Immediately following the labeling procedure,cells in suspension were treated with Antifade (MolecularProbes, Leiden, The Netherlands) and sealed under glass

coverslips. Cells were imaged using a Zeiss LSM510 confocalmicroscope (Zeiss, Welwyn Garden City, UK), using theappropriate filter sets for the fluorochromes. Excitation was at488 nm for both propidium iodide and FITC. Propidium iodidefluorescence was collected through a 570-nm dichroic mirrorand a 650-nm pass filter, and FITC fluorescence was collectedthrough a 570-nm and then a 545-nm dichroic mirror andthrough a 505–550–nm band pass filter.

RESULTS

Induction of expression of class II MHC onhealthy peripheral blood neutrophils. Freshly isolatedperipheral blood neutrophils from healthy control sub-jects did not express class II MHC molecules on theircell surface when analyzed by flow cytometry (Figure1A). However, when these cells were cultured for 20hours in the absence (Figure 1A) or presence (Figure1A) of GM-CSF and IFN�, surface expression of thisreceptor increased significantly (P � 0.01). Addition ofcytokines to the culture medium for the duration of theincubation led to only a slight increase in surface expres-sion, not significantly different from that of controlsuspensions incubated in culture medium containingFCS (Figure 1B). The addition of culture supernatantfrom peripheral blood mononuclear cells incubated withSEB (which contains a mixture of cytokines) also led toa significant increase in cell surface expression of class IIMHC on control blood neutrophils.

Cell surface expression of class II MHC onneutrophils from patients with RA. Neutrophils isolatedfrom the venous blood of patients with RA did notexpress class II MHC molecules on their cell surface.However, during incubation in culture, either in thepresence or absence of cytokines, surface expression wasup-regulated to an extent similar to that observed inblood neutrophils isolated from healthy individuals(data not shown). Similarly, neutrophils freshly isolatedfrom the SF of patients with RA did not express thisreceptor on their cell surface (Figure 2A).

To eliminate the possibility that class II MHC–expressing neutrophils within SF were lost during thepurification protocol, unfractionated SF was analyzed byimmunostaining and flow cytometry, and the neutrophilpopulation analyzed by gating on these cells by virtue oftheir forward scatter and side scatter properties. Evenusing this approach, it was not possible to detect class IIMHC expression on neutrophils within SF. However,when neutrophils isolated from SF were cultured for 20hours in the absence of any further additions (exceptFCS), surface class II MHC expression was significantlyelevated (Figure 2A). The addition of GM-CSF andIFN� to these cells did not induce expression further.

2798 CROSS ET AL

This induction of surface expression of class II MHC onSF neutrophils was more than twice that seen in bloodneutrophils after culture (Figure 1).

Unfractionated SF was also cultured for 20 hoursin the presence or absence of GM-CSF and IFN�, and

class II MHC expression on neutrophils was determinedby flow cytometry, with gating on the neutrophil popu-lation. Remarkably, surface expression on the neutro-

Figure 1. Surface expression of class II major histocompatibilitycomplex (MHC) molecules on control blood neutrophils. A, Neutro-phils were isolated from the venous blood of healthy individuals andanalyzed for surface expression of class II MHC by flow cytometry asfollows: a) time 0, second antibody only; b) time 0, first and secondantibodies; c) 20-hour culture in medium plus fetal calf serum (FCS);d) 20-hour culture in medium plus FCS supplemented withinterferon-� (IFN�) (100 units/ml) and granulocyte–macrophagecolony-stimulating factor (GM-CSF) (50 units/ml). Results shown arerepresentative of at least 20 separate experiments using blood fromdifferent donors. B, Summary data from experiments (n � 16–33)performed as described in A. 1 � time 0, second antibody only; 2 �time 0, first and second antibodies; 3 � 20-hour culture in medium plusFCS; 4 � 20-hour culture in medium plus FCS supplemented withIFN� (100 units/ml) and GM-CSF (50 units/ml); 5 � 20-hour culturein medium plus FCS supplemented with 5% conditioned medium fromStaphylococcus enterotoxin B–stimulated monocytes; 6 � 20-hourculture in medium plus FCS, second antibody only. Bars show themean and SD.

Figure 2. Surface expression of class II MHC molecules by synovialfluid neutrophils (SFNs). A, SFNs were purified as described inMaterials and Methods and incubated in the absence (control) orpresence (stimulated) of IFN� (100 units/ml) plus GM-CSF (50units/ml). SFN represents isolated neutrophils (n � 21 experiments);WSF (whole SF) represents measurements made on neutrophils (bygating) in unfractionated synovial fluid (n � 14). Values are the meanand SD at time 0 (open bars) and at 20-hour culture (filled bars). Thebroken line indicates the level of nonspecific binding. B, Representa-tive flow cytometric tracings of class II MHC expression on isolatedSFNs measured at time 0 (open tracing) or after 20 hours of incubation(shaded tracing). See Figure 1 for other definitions.

INDUCIBLE EXPRESSION OF CLASS II MHC IN INFLAMMATORY NEUTROPHILS 2799

phils was increased dramatically (Figure 2A), with levelsof expression up to 8-fold greater than those seen oncultured blood neutrophils (Figure 1B), and comparablewith those seen in professional antigen-presenting cells.This increased expression was not further enhanced bythe addition of cytokines to the incubation. The numberof class II MHC molecules on the surface of thesecultured SF neutrophils was estimated to be between30,000 and 35,000 per cell.

We also measured inducible expression of class IIMHC molecules on neutrophils isolated from the bloodand SF of patients with other inflammatory joint ar-thropathies. No endogenous or inducible expression wasdetected on neutrophils from 3 patients with psoriaticarthritis and 1 patient with systemic lupus erythemato-sus, but expression was induced on SF neutrophils froma patient with reactive arthritis (data not shown).

The potential for cultured SF neutrophils to alsoexpress costimulatory molecules CD80 and CD86 wasthen studied. However, under the same culture condi-tions that resulted in increased surface expression ofclass II MHC by either blood or SF neutrophils, surfaceexpression of CD80 and CD86 was not significantlyincreased (Figure 3), despite findings (using reversetranscriptase–polymerase chain reaction) that CD80/86messenger RNA (mRNA) is detectable in neutrophils(23). Similarly, very low levels of CD80 and CD86 weredetected after culture in the presence of GM-CSF plusIFN� (data not shown).

Determination of class II MHC expression byWestern blotting. Having used flow cytometry to estab-lish that cultured SF neutrophils expressed high levels ofclass II MHC molecules on their cell surface (Figure 2),we set out to confirm class II MHC expression using aseparate technique, namely Western blotting. The ad-vantages of this technique were 2-fold. First, identifica-tion of an immunoreacting polypeptide of the samemolecular masses as those of the � and � subunits of theclass II MHC molecule would confirm the results of flowcytometry and eliminate the possibility of nonspecificantibody binding in flow cytometry experiments. Second,although flow cytometry measured surface expression,Western blotting could detect total receptors per cell,i.e., surface receptors and newly synthesized intracellularreceptors.

Western blotting of protein extracts from freshlyisolated neutrophils from the blood of healthy individu-als or patients with RA revealed undetectable levels ofexpression of class II MHC molecules (either � or �chain), confirming the results obtained by flow cytom-

etry (Figures 4A and B). However, in contrast to theresults obtained by flow cytometry, Western blottingrevealed significantly higher protein levels in neutrophilsfreshly isolated from the SF of patients with RA (P �0.01). Because no surface expression of class II MHCmolecules was detected in these cells by flow cytometry,we conclude that in SF neutrophils from RA patients,the protein detected by Western blotting largely repre-sents intracellular or newly synthesized molecules.

These class II MHC protein levels detected byWestern blotting were increased further in control bloodneutrophils and SF neutrophils after culture in vitro withGM-CSF and IFN� (Figures 4A and B). Indeed, after20-hour culture of SF neutrophils with cytokines, totalcellular levels of class II MHC were close to thoseobserved in Raji cells.

Representative Western blots are shown in Fig-ure 4C, indicating undetectable protein levels in freshlyisolated blood neutrophils from healthy controls, butsignificant levels in freshly isolated SF neutrophils.These levels increased markedly after culture in vitro.Figure 4 also illustrates that neutrophils isolated fromthe left and right knees of a patient with RA showexactly the same changes in expression of both the � and� chain of class II MHC.

Figure 3. Expression of class II MHC and accessory molecules onneutrophils. Neutrophils were isolated from either the blood (� and■) or synovial fluid (E, F, ‚, and Œ) of patients with rheumatoidarthritis. They were then incubated in either the absence (E, �, ‚, andŒ) or presence (F and ■) of IFN� (100 units/ml) and GM-CSF (50units/ml) for 20 hours. Surface expression of class II MHC molecules(E, F, �, and ■), CD80 (‚), and CD86 (Œ) was determined by flowcytometry. Values are summary data from 5–12 experiments for eachstudy condition. See Figure 1 for definitions.

2800 CROSS ET AL

Fluorescence imaging of class II MHC expres-sion on SF neutrophils. In order to directly visualizeexpression of class II MHC molecules, isolated SFneutrophils were cultured for 20 hours and immuno-stained for receptor expression. The cells were alsostained with propidium iodide to identify the character-istic polymorphonuclear morphology of neutrophils.Figure 5A clearly shows that all of the cells in themicroscopic field had a polymorphic nucleus, which is

characteristic of nonapoptotic neutrophils. All of thesecells expressed (at varying levels) class II MHC on theircell surface, as indicated by the FITC staining (Figure5B). Class II MHC staining on neutrophils was clusteredinto discrete areas on the cell surface. Because thenuclear morphology of these cells was nonapoptotic(polymorphonuclear morphology), this class II MHCstaining was not a consequence of the neutrophils be-coming apoptotic.

Figure 4. Expression of class II major histocompatibility complex (MHC) molecules by Western blotting. Protein extracts from control blood orpurified synovial fluid (SF) and peripheral blood (PB) neutrophils from patients with rheumatoid arthritis (RA) were analyzed for expression of A,the 36-kd � subunit and B, the 29-kd � subunit of class II MHC molecules by Western blotting. Raji cells were used as a positive control. Proteinextracts were prepared at 0 hour (immediately after isolation) or after 20 hours of incubation ex vivo in the absence (�) or presence (�) ofinterferon-� (IFN�) (100 units/ml) plus granulocyte–macrophage colony-stimulating factor (GM-CSF) (50 units/ml). C, Western blot analyses ofcontrol PB neutrophils and SF neutrophils isolated from either the left knee (SFL) or right knee (SFR) of a patient with RA. Samples were analyzedeither immediately after isolation (0 hour) or after 20 hours of incubation in the absence (�) or presence (�) of IFN� (100 units/ml) plus GM-CSF(50 units/ml).

INDUCIBLE EXPRESSION OF CLASS II MHC IN INFLAMMATORY NEUTROPHILS 2801

Class II MHC expression determined by North-ern blotting. Although flow cytometry, Western blotting,and fluorescence imaging all clearly showed that SFneutrophils acquire class II MHC molecules duringculture, it was important to determine whether neutro-phils genuinely expressed this molecule de novo. Wetherefore used Northern blotting to investigate theexpression of mRNA for class II MHC molecules inneutrophils. Control blood neutrophils were incubatedfor up to 20 hours in the absence (control) or presenceof 100 units/ml IFN�, and total RNA was isolated. Theisolated RNA was then probed for transcripts for class IIMHC (HLA–DR�) and other transcripts.

Figure 6 shows that control blood neutrophilsincubated for up to 1 hour in vitro in the absence ofcytokines did not express mRNA for HLA–DR�. How-ever, after 3–5 hours of incubation with IFN�, tran-scripts for both molecules were markedly induced. Incontrast, control neutrophils expressed mRNA forp47phox, one of the cytoplasmic components of theNADPH oxidase (35), but levels of this transcript actu-ally decreased upon treatment with IFN�. MessengerRNA levels of the housekeeping gene PGK remainedconstant in both control and IFN�-treated neutrophils.Thus, IFN� treatment selectively activated the expressionof only certain genes, one of which was class II MHC.Similar results were obtained when GM-CSF and INF�were used together to stimulate expression.

We then probed mRNA isolated from SF andblood neutrophils of patients with RA to determinelevels of class II MHC transcripts in Northern blots.Figure 7 shows that significant levels of mRNA for classII MHC were detected in blood neutrophils from RA

patients (Figure 7A), but no signals were detected inneutrophils isolated from the blood of healthy controls(Figure 7B). Comparison of paired blood and SF neu-trophils from the same patient revealed that the SF

Figure 5. Fluorescence imaging of class II MHC expression by cul-tured SF neutrophils. Neutrophils were isolated from the SF of apatient with RA and cultured for 20 hours. After this incubation, theywere stained for nuclear morphology by A, propidium iodide (red), andfor surface expression of class II MHC molecules by B, indirectimmunofluorescence (green). C, Phase contrast. D, Superimposedimage of A–C. Shown are representative results from 1 of 8 experi-ments performed. See Figure 4 for definitions. Figure 6. Expression of mRNA for class II major histocompatibility

complex (MHC) molecules by peripheral blood neutrophils. Neutro-phils were isolated from the venous blood of healthy donors andincubated for up to 21 hours in the absence (control; C) or presence of100 units/ml of interferon-� (IFN�). At the indicated times, RNA wasisolated and analyzed for expression of class II MHC transcripts byNorthern blotting. Filters were then stripped and consecutively probedfor expression of p47phox and the housekeeping gene PGK. Shown arerepresentative results from 1 of 6 experiments performed, eachinvolving a different donor.

Figure 7. Expression of mRNA for class II major histocompatibilitycomplex (MHC) molecules by synovial fluid (SF) neutrophils. TotalRNA was isolated from the peripheral blood (PB) of 3 healthy controlsand from the PB and SF of 3 patients with rheumatoid arthritis (RA).Total RNA was also isolated from Raji cells used as a positive control.RNA was then analyzed for expression of class II MHC transcripts byNorthern blotting.

2802 CROSS ET AL

neutrophils consistently expressed higher levels ofmRNA for class II MHC compared with levels presentin blood neutrophils.

T cell proliferation by class II MHC–expressingneutrophils. Neutrophils were isolated from the bloodand SF of patients with RA and then incubated in vitroto induce expression of class II MHC molecules. Theywere then incubated with superantigen and then withCD4� T cells. Monocytes treated and incubated underidentical conditions were able to stimulate T cell prolif-eration, as detected by tritiated thymidine incorporation(Figure 8). SF neutrophils treated under identical con-ditions could also stimulate T cell proliferation, theiractivity being a mean � SD of 23 � 10% that ofmonocytes (n � 3 measurements). Thus, class II MHC–expressing neutrophils are capable of stimulating T cellproliferation. Control experiments using a blocking an-tibody (Figure 9) confirmed that this T cell proliferationwas attributable to class II MHC molecules and not to asoluble factor released by the neutrophils.

DISCUSSION

This is the first study to show that SF neutrophilsfrom patients with RA can transcribe, synthesize, andexpress high levels of class II MHC molecules on theircell surface. In contrast to other reports detailing detec-tion of class II MHC molecules on the cell surface ofneutrophils (21,23), we have shown not only that thesemolecules are synthesized by neutrophils, but also thatthe SF environment is sufficient for such expression,without the requirement for any exogenous agents suchas cytokines. Surface expression was further enhancedwhen neutrophils were incubated in unfractionated SF,but also occurred (albeit at lower levels) when purifiedSF neutrophils were cultured in RPMI medium supple-mented with FCS. Taken together, these observationsindicate that induction of class II MHC molecules inneutrophils is triggered within the SF of rheumatoidjoints, and that SF factors can further enhance thisexpression during ex vivo culture.

If class II MHC expression is indeed triggered inneutrophils within rheumatoid SF, then it is curious thatno surface class II MHC expression was detected onfreshly isolated SF neutrophils. Surface expression wasdetected only after ex vivo culture, with or without SF.However, although freshly isolated SF neutrophils hadno detectable surface expression of class II MHC, weconsistently detected positive signals for this molecule byWestern blotting. We also consistently detected mRNAfor class II MHC by Northern blotting. This mRNA waspresent at low but detectable levels in freshly isolatedblood neutrophils from RA patients but was not de-tected at all in freshly isolated neutrophils from healthy

Figure 8. T cell proliferation by synovial fluid (SF) neutrophils.Neutrophils were isolated from the SF of patients with rheumatoidarthritis and cultured for 24 hours to induce expression of class IImajor histocompatibility complex molecules on their cell surface. Theywere then incubated with superantigen and then with CD4� T cells for24 hours, as described in Materials and Methods. T cell proliferationwas then determined by incorporation into tritiated thymidine. M �blood mononuclear phagocytes; T � CD4� T cells; SF-PMN �synovial fluid polymorphonuclear neutrophils. Values are the meanand SD of 5 replicate measurements.

Figure 9. Effect of blocking antibodies on T cell proliferation. Neu-trophils were isolated from the synovial fluid (SF) and blood ofpatients with rheumatoid arthritis and cultured for 24 hours to induceexpression of class II major histocompatibility complex molecules ontheir cell surface. They were then incubated with superantigen andthen with CD4� T cells for 48 hours, as described in Materials andMethods. Cultures were also incubated with (�) or without (�) L243antibody, and its effects on T cell proliferation were determined.Monocytes � mononuclear phagocytes; PBN � peripheral bloodneutrophils; SFN � SF neutrophils. Shown are the combined datafrom 4 separate experiments; bars show the mean and SD of 4 replicatemeasurements.

INDUCIBLE EXPRESSION OF CLASS II MHC IN INFLAMMATORY NEUTROPHILS 2803

controls. Transcripts for class II MHC were also de-tected in greater amounts in freshly isolated SF neutro-phils.

Therefore, it may be concluded that in the pop-ulation of SF neutrophils, expression of class II MHCmolecules has been triggered, so that mRNA for thisreceptor is transcribed, and low levels of this mRNA aretranslated. However, this newly synthesized protein ap-pears to be intracellular, representing molecules thatmay not have been fully processed and assembled on thecell surface. Indeed, because peripheral blood neutro-phils from RA patients expressed detectable mRNA forclass II MHC, it is likely that the initial triggering ofexpression in RA occurs in the blood of these patients.Messenger RNA levels then increase further when thecells enter the SF, and the protein becomes expressedbut remains intracellular. The relative levels of tran-scripts, intracellular protein, and cell surface class IIMHC are shown in Table 1. Transcripts for class IIMHC actually decreased after culture of SF neutrophilsin vitro (data not shown), indicating that the majorprocess that occurs during culture is the translation ofexisting mRNA molecules and the processing of thesemolecules to become expressed on the cell surface.

If this hypothesis is correct, then it is interestingto ask: what happens to the SF neutrophils once theyexpress class II MHC molecules on their cell surface invivo? Clearly, they cannot be detected in freshly isolatedSF. We first thought that perhaps these class II MHC–expressing neutrophils were fragile and were not recov-ered during our cell isolation and purification proce-dures. However, when we analyzed unfractionated SF,we still could not detect surface expression by flowcytometry, after gating on the neutrophil population. Itis thus possible that once SF neutrophils express class IIMHC on their cell surface within the joint, then theymay be removed from the fluid. This may occur if theneutrophils are phagocytosed by, for example, macro-

phages, but we did not observe this when neutrophilswere coincubated with macrophages during ex vivoculture of unfractionated SF. Alternatively, class IIMHC–expressing SF neutrophils may move from thefluid phase to a solid phase, such as the synovialmembrane or cartilage. It will be extremely important toidentify the fate of class II MHC–expressing SF neutro-phils in order to elucidate the biologic function of thesecells. Neutrophils have been detected, sometimes inlarge numbers, at the cartilage–pannus interface (36). Itwill now be necessary to determine whether these neu-trophils are class II MHC positive.

Expression of class II MHC molecules was de-pendent on de novo transcription and translation.Clearly, this activation of gene expression was selective,because expression of class II MHC was markedlyup-regulated by IFN�, whereas expression of the house-keeping gene PGK was unaltered, and expression ofp47phox was down-regulated. However, there is a cleardiscrepancy between the levels of induction of class IIMHC mRNA levels and surface expression of thismolecule following IFN� treatment. One possible expla-nation is that another signal is required to translateand/or process the transcript into a functional cellsurface molecule. This second signal is likely to bepresent in rheumatoid SF. It is also necessary to deter-mine what other neutrophil gene products are up-regulated (and down-regulated) in SF neutrophils. Iden-tification of newly expressed proteins that are notexpressed by blood neutrophils will shed further insightsinto the function of SF neutrophils and their role indisease pathology. The detection of elevated levels ofmRNA for class II MHC molecules in neutrophils is animportant observation, because it unambiguously showsthat expression of this molecule by neutrophils requiresdo novo biosynthesis. This finding eliminates the possi-bility that neutrophils have phagocytosed and subse-quently expressed on their cell surface class II MHC

Table 1. Relative levels of mRNA, intracellular protein, and surface protein of class II MHC moleculesin RA neutrophils*

PB neutrophils,freshly isolated

SF neutrophils,freshly isolated

SF neutrophils,after culture

Messenger RNA � ��� �Intracellular protein, newly synthesized �/� �� ���Surface protein � � ���

* Neutrophils were isolated from the peripheral blood (PB) or synovial fluid (SF) of patients withrheumatoid arthritis (RA) and analyzed either immediately after isolation or after culture for 20 hours.Messenger RNA was measured by Northern blotting, intracellular protein was measured by Westernblotting, and surface protein was measured by flow cytometry. MHC � major histocompatibility complex.

2804 CROSS ET AL

molecules that have been shed from cells such as mono-cytes.

It has recently been reported that neutrophilsisolated from the cerebrospinal fluid of some patientswith Wegener’s granulomatosis can also express class IIMHC on their cell surface (37), but neither intracellularmolecules nor mRNA was investigated. Our finding ofclass II MHC on SF neutrophils is, to our knowledge, thefirst indication that this phenomenon may be involved inthe immune responses that occur in RA. It is thusimportant to fully understand the function and hence thesignificance of this phenomenon. The only known func-tion of class II MHC is to present antigen to CD4� Tlymphocytes, and indeed this may be the function of thisreceptor in neutrophils. Although under the experimen-tal conditions described in this report we failed to detectsignificant expression of the costimulatory moleculesCD80 and CD86 on neutrophils that expressed highlevels of class II MHC, these cells were still capable ofstimulating T cell proliferation (Figure 8). Further workis necessary to establish the function and role of class IIMHC in SF neutrophils, specifically in RA, but moregenerally in the regulation of T cell biology. Neutrophilscould thus play an as yet unrecognized role in theregulation of immune responses.

REFERENCES

1. De Keyser F, Elewaut D, Vermeersch J, De Wever N, Cuvelier C,Veys EM. The role of T cells in rheumatoid arthritis. ClinRheumatol 1995;14 Suppl 2:5–9.

2. Choy EH, Kingsley GH, Panayi GS. Innovative treatment ap-proaches for rheumatoid arthritis: T-cell regulation. Baillieres ClinRheumatol 1995;9:653–71.

3. Weyand CM, Goronzy JJ. Pathogenesis of rheumatoid arthritis.Med Clin North Am 1997;81:29–55.

4. Cantagrel A, Lambert N, Alam A. T cell receptor gene in synovialtissues of rheumatoid arthritis. Int Rev Immunol 1998;17:323–37.

5. Weyand CM, Bryl E, Goronzy JJ. The role of T cells in rheuma-toid arthritis. Arch Immunol Ther Exp (Warsz) 2000;48:429–35.

6. Dolhain RJ, Ter Haar NT, De Kuiper R, Nieuwenhuis IG,Zwinderman AH, Breedveld FC, et al. Distribution of T cells andsigns of T-cell activation in the rheumatoid joint: implications forsemiquantitative comparative histology. Br J Rheumatol 1998;37:324–30.

7. Richardson C, Emery P. Clinical use of cyclosporin in rheumatoidarthritis. Drugs 1995;50:26–36.

8. Pillinger MH, Abramson SB. The neutrophil in rheumatoid arthri-tis. Rheum Dis Clin North Am 1995;21:691–714.

9. Edwards SW, Hallett MB. Seeing the wood for the trees: theforgotten role of neutrophils in rheumatoid arthritis. ImmunolToday 1997;18:320–4.

10. Dahlgren C, Karlsson A. Respiratory burst in human neutrophils.J Immunol Methods 1999;232:3–14.

11. Kitsis E, Weissmann G. The role of the neutrophil in rheumatoidarthritis. Clin Orthop 1991:63–72.

12. Wipke BT, Allen PM. Essential role of neutrophils in the initiation

and progression of a murine model of rheumatoid arthritis.J Immunol 2001;167:1601–8.

13. Moulding DA, Quayle JA, Hart CA, Edwards SW. Mcl-1 expres-sion in human neutrophils: regulation by cytokines and correlationwith cell survival. Blood 1998;92:2495–502.

14. Robinson J, Watson F, Bucknall RC, Edwards SW. Activation ofneutrophil-reactive oxidant production by synovial fluid frompatients with inflammatory joint disease: soluble and insolubleimmunoglobulin aggregates activate different pathways in primedand unprimed cells. Biochem J 1992;286:345–51.

15. Haslett C. Granulocyte apoptosis and inflammatory disease. BrMed Bull 1997;53:669–83.

16. Quayle JA, Watson F, Bucknall RC, Edwards SW. Expression ofFc gamma RIII in neutrophils in rheumatoid arthritis. BiochemSoc Trans 1996;24:489S.

17. Cassatella MA. The production of cytokines by polymorphonu-clear neutrophils. Immunol Today 1995;16:21–6.

18. Condliffe AM, Kitchen E, Chilvers ER. Neutrophil priming:pathophysiological consequences and underlying mechanisms.Clin Sci (Colch) 1998;94:461–71.

19. Feldmann M, Brennan FM, Maini R. Cytokines in autoimmunedisorders. Int Rev Immunol 1998;17:217–28.

20. Quayle JA, Watson F, Bucknall RC, Edwards SW. Neutrophilsfrom the synovial fluid of patients with rheumatoid arthritisexpress the high affinity immunoglobulin G receptor, Fc gammaRI (CD64): role of immune complexes and cytokines in inductionof receptor expression. Immunology 1997;91:266–73.

21. Gosselin EJ, Wardwell K, Rigby WF, Guyre PM. Induction ofMHC class II on human polymorphonuclear neutrophils by gran-ulocyte/macrophage colony-stimulating factor, IFN-gamma, andIL-3. J Immunol 1993;151:1482–90.

22. Oehler L, Majdic O, Pickl WF, Stockl J, Riedl E, Drach J, et al.Neutrophil granulocyte-committed cells can be driven to acquiredendritic cell characteristics. J Exp Med 1998;187:1019–28.

23. Radsak M, Iking-Konert C, Stegmaier S, Andrassy K, Hansch GM.Polymorphonuclear neutrophils as accessory cells for T-cell acti-vation: major histocompatibility complex class II restricted anti-gen-dependent induction of T-cell proliferation. Immunology2000;101:521–30.

24. Cush JJ, Lipsky PE. Cellular basis for rheumatoid inflammation.Clin Orthop 1991;Apr:9–22.

25. Xu WD, Firestein GS, Taetle R, Kaushansky K, Zvaifler NJ.Cytokines in chronic inflammatory arthritis. II. Granulocyte-macrophage colony-stimulating factor in rheumatoid synovial ef-fusions. J Clin Invest 1989;83:876–82.

26. Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF,Cooper NS, et al. The American Rheumatism Association 1987revised criteria for the classification of rheumatoid arthritis.Arthritis Rheum 1988;31:315–24.

27. Edwards SW. The O-2 generating NADPH oxidase of phagocytes:structure and methods of detection. Methods 1996;9:563–77.

28. Boyum A. A one-stage procedure for isolation of granulocytes andlymphocytes from human blood: general sedimentation propertiesof white blood cells in a 1g gravity field. Scand J Clin Lab InvestSuppl 1968;97:51–76.

29. Hamel ME, Eynon EE, Savelkoul HF, van Oudenaren A, Kruis-beek AM. Activation and re-activation potential of T cells re-sponding to staphylococcal enterotoxin B. Int Immunol 1995;7:1065–77.

30. Jappe U. Superantigens and their association with dermatologicalinflammatory diseases: facts and hypotheses. Acta Derm Venereol2000;80:321–8.

31. Shapiro HM. Practical flow cytometry. 2nd ed. New York: Alan R.Liss; 1988. p. 163–4.

32. Laemmli UK. Cleavage of structural proteins during the assemblyof the head of bacteriophage T4. Nature 1970;227:680–5.

INDUCIBLE EXPRESSION OF CLASS II MHC IN INFLAMMATORY NEUTROPHILS 2805

33. Piccinini LA, Schachter BS, Davies TF. HLA-DR alpha chainexpression in human thyroid cells. Endocrinology 1986;118:2611–3.

34. Chadderton T, Wilson C, Bewick M, Gluck S. Evaluation of threerapid RNA extraction reagents: relevance for use in RT-PCR’sand measurement of low level gene expression in clinical samples.Cell Mol Biol (Noisy-le-grand) 1997;43:1227–34.

35. Cassatella MA, Bazzoni F, Flynn RM, Dusi S, Trinchieri G, RossiF. Molecular basis of interferon-gamma and lipopolysaccharideenhancement of phagocyte respiratory burst capability: studies on

the gene expression of several NADPH oxidase components. JBiol Chem 1990;265:20241–6.

36. Mohr W, Menninger H. Polymorphonuclear granulocytes at thepannus–cartilage junction in rheumatoid arthritis. ArthritisRheum 1980;23:1413–4.

37. Hansch GM, Radsak M, Wagner C, Reis B, Koch A, Breitbart A,et al. Expression of major histocompatibility class II antigens onpolymorphonuclear neutrophils in patients with Wegener’s gran-ulomatosis. Kidney Int 1999;55:1811–8.

2806 CROSS ET AL