Research paper

Rational design of synthetic peptides to generate antibodies thatrecognize in situ CD11c+ putative dendritic cells in horse lymph nodes

Gerardo P. Espino-Solis a, J. Calderon-Amador b, E.S. Calderon-Aranda c,A.F. Licea d, L. Donis-Maturano b, L. Flores-Romo b, L.D. Possani a,*a Departamento de Medicina Molecular y Bioprocesos, Instituto de Biotecnologıa, Universidad Nacional Autonoma de Mexico, 62210 Morelos, Mexicob Departamento de Biologıa Celular, CINVESTAV-IPN, Mexico, D.F., Mexicoc Toxicologıa, CINVESTAV-IPN, Mexico, D.F., Mexicod Departamento de Biotecnologia Marina, CICESE, Ensenada, B.C., Mexico

Veterinary Immunology and Immunopathology 132 (2009) 181–190

A R T I C L E I N F O

Article history:

Received 12 December 2008

Received in revised form 8 May 2009

Accepted 10 June 2009

Keywords:

Antibody

CD11c

Dendritic cells

Immunohistochemistry

Horse

Lymph node

Synthetic peptides

A B S T R A C T

A three-dimensional model of the aX I-domain of the horse integrin CD11c from dendritic

cells provided information for selecting two segments of the primary structure for peptide

synthesis. Peptide 1 contains 20 amino acids and peptide 2 has 17 amino acid residues. The

first spans from position Thr229 to Arg248 of an a-helix segment of the structure, whereas

peptide 2 goes from Asp158 to Phe174 and corresponds to an exposed segment of the loop

considered to be the metal ion-dependent adhesion site. Murine polyclonal antisera

against both peptides were generated and assayed in peripheral blood cell suspensions

and in cryosections of horse lymph nodes. Only the serum against peptide 2 was capable of

identifying cells in suspension and in situ by immunohistochemistry, some with evident

dendritic morphology. Using this approach, an immunogenic epitope exposed in CD11c

was identified in cells from horse lymph node in situ.

� 2009 Elsevier B.V. All rights reserved.

Contents lists available at ScienceDirect

Veterinary Immunology and Immunopathology

journal homepage: www.e lsev ier .com/ locate /vet imm

1. Introduction

A distinctive feature of the immune system is itscapacity to recognize and respond to a great diversity offoreign molecules called antigens. The proper function ofthe system involves a dynamic and complex interactionamong different types of cells and soluble mediators calledcytokines (Banchereau et al., 2000; Mellman, 2007).Dendritic cells (DCs) play a crucial role in the immuneresponse due to their strategic location in peripheraltissues and within lymphoid organs (Flores-Langaricaet al., 2005; Flores-Romo, 2001; Jimenez-Flores et al.,2006) and also because their excellent capacity to captureand process antigens, and present antigen-derived pep-tides in the context of MHC molecules to naıve T cells andcommit suitable co-estimulatory signals that command

* Corresponding author.

E-mail address: possani@ibt.unam.mx (L.D. Possani).

0165-2427/$ – see front matter � 2009 Elsevier B.V. All rights reserved.

doi:10.1016/j.vetimm.2009.06.017

either immunogenic or tolerogenic T cell stimulation. Theencounter of DCs with antigens in peripheral tissues allowschanneling the complex to the lymphoid organs, wherethey can make contact with T cells and initiate the immuneresponse (Banchereau et al., 2000; Banchereau and Stein-man, 1998). To accomplish these multiple tasks, DCs areequipped with an advanced endocytic system and a widerepertoire of surface receptors such as adhesion moleculesand integrins (Ueno et al., 2007). Integrins are adhesionreceptors consisting of non-covalently associated a and bsubunits with molecular weights of 120–180 kDa and 90–110 kDa, respectively. Integrins can mediate cell–cell, cell–extra-cellular matrix and cell–pathogen interactions. Theyparticipate in critical roles of the immune system duringleukocyte traffic and migration, in the formation of theimmunological synapse, and in co-stimulation and pha-gocytosis (Hynes, 2002). In vertebrates, integrins consti-tute a family of transmembrane adhesion receptorscomprised by 18 different a subunits and 8 different bsubunits which assemble in 24 known a/b heterodimers

G.P. Espino-Solis et al. / Veterinary Immunology and Immunopathology 132 (2009) 181–190182

(Luo et al., 2007). Half of integrin a subunits contain adomain of about 200 amino acids known as inserted-domain (I-domain), or von Willebrand factor A-domain(vWA). In integrins where it is present, the aI-domain isthe major or exclusive ligand-binding site. The aI-domaincan be expressed independently of other integrin domainsand was the first domain crystallized. The folding of thisdomain resembles that of small G proteins, with sevenamphipathic helices surrounding a hydrophobic b-sheetcore. One Mg2+ ion is linked at a metal ion-dependentadhesion site (MIDAS) at the ligand-binding ‘‘top’’ face ofthe domain, at the C-terminal end of the parallel b-strands.Several structures of aI-domains are now available,including aX, aM, aL and a2 (Luo et al., 2007; Vorup-Jensen et al., 2003). The integrin b2 subfamily isexclusively expressed on leukocytes, and is composed ofa unique a subunit (CD11a, CD11b, CD11c, and CD11d)complexed with a common b subunit (CD18). The integrinCD11c/CD18 (aXb2, p150/95, complement receptor 4)(Luo et al., 2007), is expressed in dendritic cells, mono-cytes, macrophages, natural killer cells and also on certainactivated T and B cells (Bilsland et al., 1994). CD11c is adendritic cell-restricted molecule expressed by mouse DCsand human DCs of myeloid origin (Pulendran et al., 2001).Usually, high levels of CD11c are found in immature DCs,which show the best efficiency for antigen capture inperipheral tissues (Banchereau and Steinman, 1998).

Horses have played an important role in manydifferent stages of human civilization, and have beenused for different tasks, with positive impact inhuman health (Steinbach et al., 2002a). Recently, thescientific research using the horse as a model animal hasadvanced in a significant manner. It includes cloning andexpression of horse cytokines, proteomic characteriza-tion of horse serum, in vitro culture of horse dendriticcells and even a project for the determination of thehorse genome is currently underway (Mauel et al., 2006;Steinbach et al., 2002b). However, equine immunologistslack the great diversity of immunological tools thatresearchers in human or mouse immunology have, suchas monoclonal antibodies (mAbs) directed againstdifferent cells, particularly to a great variety of cellsurface molecules (antigen receptors, cytokines, anti-bodies, complement, adhesion molecules and integrins,among others) (Ibrahim and Steinbach, 2007). Many ofthese molecules are poorly known or not characterizedat all in equines.

This work is an additional step of a long-term projectwith significant biotechnological and immunologicalimpact. Our main objective is the production of morespecific antibodies in a shorter time, through the use ofdendritic cells. The initial effort aims at the production ofhorse antibodies. For this purpose we first obtained thecomplete sequence of aX subunit of the horse integrinCD11c. Then it was possible to construct a three-dimensional (3D) model of the I-domain of the aX subunitof the horse CD11c, based on the fold prediction analysis ofthe horse CD11c integrin (Espino-Solis et al., 2008)prepared with the information generated with the knownstructure of the I-domain from the human aX (Vorup-Jensen et al., 2003). Analysis of the above mentioned 3D

model, allowed us to select a pair of segments of the entiremolecule for peptide synthesis. The rationale was toidentify specific epitopes of the I-domain of the aX subunitof CD11c capable of inducing the production of murinepolyclonal antibodies, but more importantly, to determineif they are exposed on the native integrin. This wouldfacilitate recognition of dendritic cells in horse lymphoidtissues by these polyclonal antibodies. To the best of ourknowledge this is the first time that this methodologicalapproach is applied to horses, allowing direct identifica-tion of an immunogenic epitope, exposed in the CD11c ofputative dendritic cells within lymph nodes, which is thesubject of this communication.

2. Material and methods

2.1. Fold recognition analysis

Secondary structure predictions were performed usingseveral programs from the Web servers, among which is:three-dimensional position-specific scoring matrix (3D-PSSM), structural classification of proteins (SCOP), pre-dictProtein (PHD), PHYRE, PROSITE and others, for detailssee Espino-Solis et al. (2008). Upon request, PHD alsoreturns fold recognition results obtained by prediction-based threading, CHOP domain assignments, predictions oftransmembrane strands and inter-residue contacts are alsoavailable (Rost et al., 2004).

Another program is 123D+ which combines sequenceprofiles, secondary structure prediction, and contact capa-city potentials to thread a protein sequence through a set ofstructures. Contact capacity potentials reflect mainlyhydrophobicity of the amino acids. Without pairwisecontact potentials the program uses a simple and fastdynamic programming algorithm to align a sequence with astructure (Alexandrov et al., 1996). These programs (3D-PSSM, PHYRE, PHD and 123D+) are available to thecommunity at the web servers: http://www.bmm.icne-t.uk/servers/3dpssm, www.sbg.bio.ic.ac.uk/�phyre/, www.predictprotein.org and http://123d.ncifcrf.gov/123D+.html.

2.2. 3D-structure model

The sequence of horse I-domain was compared againstProtein Data Bank (PDB) with BLAST protocol, aiming atthe identification of the most suitable structure forhomology modeling. A computational model of thethree-dimensional structure of horse I-domain (accessionnumbers NM_001114177.1, EU_196238) was built by theprogram Swiss Model Repository Server (http://swissmo-del.expasy. org/repository/), using as template the crys-tallographic structures of I-domains from the integrins aXand aL (PDB entry 1N3y and 1jlm, respectively) (Arnoldet al., 2006).

2.3. Rationale choice of two peptides for chemical synthesis

The 3D-structure modeling of I-domain allowed select-ing two segments of amino acid sequences of the recentlycloned integrin for chemical synthesis. They were calledpeptide 1, which comprehends the sequence that goes

G.P. Espino-Solis et al. / Veterinary Immunology and Immunopathology 132 (2009) 181–190 183

from Thr229 to Arg248; and peptide 2, corresponding tothe amino acid sequence from position Asp158 to Phe174of the full amino acid sequence of the integrin CD11c. Thefirst rationale was to choose segments of the primarystructure that show low sequence identity with otherknown integrins (see below). In addition, the choice wasmade by visual inspection of the model, in such a way thattwo differently folded segments of the I-domain would berepresented by the synthetic peptides; one correspondingto a segment of the exposed a-helix and the othercorresponding to the loop of the metal ion-dependentadhesion site (MIDAS), assumed to be the face by which theintegrin binds their ligands. The idea was to select first, abona fide stable a-helix folded segment (peptide 1); andsecond, a distinct folded segment, more exposed to thesurface of the I-domain (peptide 2). The hypothesis aimedat obtaining at least one of these segments, capable ofinducing antibodies that would recognize the nativeprotein in a rather specific manner. Although the choicecan be regarded as arbitrary, as it will become clear in thediscussion below, this strategy was successful, sinceantibodies that recognize the I-domain were obtained byimmunizing with these synthetic peptides.

2.4. Preparation of synthetic peptides

Peptide synthesis was accomplished by the solid phasemethod of Merrifield, using fluorenylmethyloxycarbonyl(Fmoc) amino acids (Nova Biochem, Darmstadt, DE)(Sheppard, 1989). The entire synthesis was performedmanually. The Rink Amide Resin (Nova Biochem, Darm-stadt, DE) was used to produce amidated C-terminalpeptides. After cleaving the peptides from the resins bymeans of ‘‘Reagent K’’ (Sheppard, 1989), the syntheticpeptides were purified by high performance chromato-graphy (HPLC) as described below.

2.5. Chemical characterization of synthetic peptides

Synthetic peptides (1 mg) were dissolved separatelyin 0.1% aqueous trifluoracetic acid (TFA), and theinsoluble material was removed by centrifugation at14,000 � g for 5 min. The supernatant was used directlyfor HPLC purification. The diluted peptides were fractio-nated using a reverse-phase semipreparative C4 columnVydac (Deerfield, IL, USA) equilibrated in 0.1% TFA, andeluted with a linear gradient run from solution A (0.1%TFA in water) to 60% solution B (0.1% TFA in acetonitrile),during 60 min at a flow rate of 2 mL/min, as earlierdescribed (Calderon-Aranda et al., 1999). Effluent absor-bance was monitored at 230 nm. Fractions were collectedin 1.5 mL borosilicate tubes and dried out in a SpeedVac,model SC110 apparatus (Savant Instruments, Inc. Farm-ingdale, NY, USA). The primary structure of the peptideswas analyzed by direct sequencing (Edman degradation),using a LF3000 Protein Sequencer (Beckman, CA, USA)and techniques already described by our group (Cal-deron-Aranda et al., 1999). The theoretical mass wasconfirmed by electrospray ionization mass spectrometrymeasurements using a Finnigan LCQDUO ion trap massspectrometer (San Jose, CA, USA).

2.6. Immunization protocol

Two groups of five female mice (strain BALB/c) ofapproximately 20 g body weight each were used at thestarting point of the immunization protocol. The mice ofeach group were separately bled to obtain the preimmuneserum before the first immunization with each one of thepeptides. The animals were immunized four times atintervals of 10 days prior bleeding for obtaining theantiserum. The synthetic peptides were used at 50 mg perdose with Freund’s adjuvant the first injection and withincomplete adjuvant for subsequent applications. For thefirst injection the peptides were coupled to tyroglobulin, asearlier described by our group (Calderon-Aranda et al.,1995), whereas the subsequent immunization were donewith peptides alone. Immunization and handle proceduresfor these animals are all authorized by the EthicalCommittee of Cinvestav Animal Facilities (CICUAL) forthe correct use of animals in experimentation (codenumber 302/06).

2.7. Titration of antisera to the synthetic peptides

Serum samples obtained from preimmunized mice(negative control) and after the immunization protocolwith synthetic peptides were all assayed on ELISA platescoated with the synthetic peptides used for immunization.Plates were coated with 5 mg/mL of peptides in 20 mMsodium bicarbonate buffer, pH 9.4 for 2 h at 37 8C.Subsequently, wells were blocked with 1% bovine serumalbumin (BSA) in phosphate buffered saline (PBS) during1 h at 37 8C and washed four times. After 2 h incubationwith serial dilutions of antiserum at 37 8C, the wells werewashed with the washing solution PBS Tween-20 0.01%.Binding of primary antibody was measured with asecondary antibody, goat anti-mouse IgG, coupled tohorseradish peroxidase (dilution 1:2000). The titer of thesera (dilution corresponding to 50% of the maximumabsorption) was estimated by linear regression.

2.8. Immunolabelling of blood cell suspensions and horse

lymph nodes in situ using murine antibodies against CD11c

synthetic peptides

To perform a preliminary assessment with antibodies topeptide 2 and the corresponding preimmune serum byflow cytometry (FACS), cell suspensions from peripheralblood of a healthy horse were used after ficoll separationand culture. Primary antibodies were used at a dilution1:500 followed by secondary antibodies goat anti-mouseconjugated to Cy5 (Invitrogen, Carlsbad, CA, USA) diluted1:800 in the flow cytometry kit solution (Becton Dickinson,San Jose, CA, USA). At least 200,000–300,000 events wererecorded for each of the samples evaluated. For theimmunohistochemical in situ studies, a lymph node (LN)from the sub-maxillary area was obtained from a healthyhorse and placed in cold sterile endotoxin-free salinesolution. The LN was embedded in Jung’s tissue freezingmedium (Leica instruments, Wetzlar, DE.), immediatelyfrozen in liquid nitrogen and stored at �80 8C until use. LNwas cut in 6 mm frozen sections using a cryostat (Leica

G.P. Espino-Solis et al. / Veterinary Immunology and Immunopathology 132 (2009) 181–190184

1900 CM, Wetzlar, DE) and fixed in cold acetone for 20 min.To block endogenous peroxidase activity the tissuesections were treated with 9% H2O2 in phosphate bufferedsaline (PBS) for 1 h at room temperature, and washed inPBS containing 0.1% BSA. Before specific antibody (Ab)staining, non-specific binding was blocked with ‘‘universalblocking reagent’’ (Power Block, Biogenex, San Ramon, CA,USA) 20 min at room temperature. Trying to detect CD11c+

DCs in situ, 100 ml of PBS containing either the preimmuneserum (isotype control) or the hyper-immune serum (bothdiluted 1:250) was added to tissue sections. Preimmuneserum at dilution 1:250 was used as negative control. Afterover night incubation at 4 8C, all samples were washedtwice with PBS/0.2% BSA and incubated with biotinylatedgoat anti-mouse antibody (Vector, Burlingame, CA, USA) inPBS/BSA 1% for 1 h at room temperature, at a dilution1:1000. After the corresponding washings, streptavidin(SAV) coupled to horseradish peroxidase was incubated for20 min at room temperature and following three washes,the chromogen (blue-grey) substrate SG (Vector, Burlin-game, CA, USA) was added and the reaction was developedat room temperature. Slides were washed extensively withdistilled water and counterstained with methyl green(Dako Carpinteria, CA, USA), rinsed with distilled water andthe slides were mounted in 60% glycerol PBS and analyzedunder a light microscope (Nikon Corp., Japan).

2.9. Screening for in situ crossreactions of antibodies to

human DC in horse lymph node cryosections

The following anti-human antibodies were used tryingto assess potential crossreactivities with cells from horselymph node: anti-MHC-II (DR) and anti-CD1a from (Dako,Carpinteria, CA, USA) both diluted 1:50; anti-CD86 andanti-CD11c from (BD-Pharmingen, San Jose, CA, USA)diluted 1:50. For the germinal centre histochemistry, thehorseradish peroxidase-conjugated peanut agglutinin(HRP-PNA) lectin from Arachis hypogaea (Vector Burlin-game, CA, USA) was used at 1:2000 dilution. Afterwashings, PNA binding was revealed with a brown-redprecipitate using VIP chromogen (Vector, Burlingame, CA,USA), rinsed with distilled water and counterstained withmethyl green (Dako, Carpinteria, CA, USA). Slides werethen coversliped and analyzed under a light microscope(Nikon Corp., Japan). In serial sections, a hematoxylin–eosin (HE) staining was performed to evidence the basichistology features.

Table 1

Best score results from fold prediction analysis.

Program PDB Fold Domain

PHYRE 1mf7a vWA-like n.d.

1n3ya vWA-like n.d.

3D-PSSM 1n3ya Integrin A (or I) domain n.d.

1ido Integrin A (or I) domain n.d.

PHD/Prodom 1n3y von Willebrand factor, type A integrin alph

1ao3 vWA-like von Willebra

123+/SCOP 1ido vWA-like von Willebra

For the prediction reported here the sequences that comprehend the primary st

3. Results

3.1. Fold prediction analysis and three-dimensional model of

horse aX I-domain

To validate the identification of the horse aX I-domainsequence, fold prediction programs were used as describedin Section 2. The amino acid sequence between residues137–348 of the full sequence from horse CD11c integrinwere submitted to the servers. The results obtained afterapplication of the different fold prediction programs alwaysqualified our sequence with best scores and showed that thefold was similar to the integrin I-domain, or von Willebrand,fold of known integrins from the literature (see Table 1). Thethree-dimensional model of horse aX I-domain was built bySwiss-Model Repository server, using as template thecrystallographic structures of aM and aX (PDB entries:1jlm and 1N3y) which has sequence identity of 74% and 57%,respectively, with the horse aX I-domain. The structure wasanalyzed with MacPyMol program. The result of the three-dimensional model of the horse I-domain adopts a a/bRossman fold, characterized by alternating seven amphi-pathic helices surrounding a hydrophobic b-sheet core. It isworth mentioning that the model obtained showed theMIDAS (metal ion-dependent adhesion site) at the top faceof the domain, as expected (Fig. 1A).

3.2. Preparation and chemical characterization of synthetic

peptides

Based on the model of the horse aX I-domain (Fig. 1A),two segments of sequence were chosen taking intoconsideration the parameters mentioned above, and thecorresponding synthetic peptides were manually preparedusing the Fmoc chemistry, as described in Section 2. Thepeptide 1 contains 20 amino acids with the sequenceTASAIQVVIKELFSATRGAR, and peptide 2 contains 17amino acids with the sequence: DGSGSIYFKDFAKMLSF.The results of their purification by HPLC are shown inFig. 1B and C, respectively. Asterisks indicate the elutiontime of pure peptides. They were sequenced by Edmandegradation showing to contain the expected amino acidsequences. Additionally, the molecular weight of the twopeptides were determined by mass spectrometry andshown to be: 2118 atomic mass units (a.m.u.) for peptide 1,and 1912 a.m.u. for peptide 2, confirming the expectedvalues. These peptides have an identity under 65% (peptide

Family Superfamily

Integrin A (or I) domain vWA-like

Integrin A (or I) domain vWA-like

Integrin A (or I) domain Integrin A (or I) domain

Integrin A (or I) domain Integrin A (or I) domain

a-X beta2 Integrin A (or I) domain vWA-like

nd factor A3 Integrin A (or I) domain vWA-like

nd factor A3 Integrin A (or I) domain vWA-like

ructure of integrin CD11c from residues numbers 137–348 were chosen.

Fig. 1. Three-dimensional model ofhorse I-domain. (A) In green color is the segmentcorresponding topeptide 1 and inred color the position ofpeptide 2. (Band C)

Chromatographic profiles of peptide purification: asterisk in B shows elution time ofpeptide 1, in C thatof peptide 2. (D and E) Multiple alignment of the segments

corresponding to the amino acid sequences of peptide 1 and peptide 2 with CD11c integrins of other species. Different amino acid residues are highlighted. The

accession numbers corresponding to the various integrins are included, for Ec, horse; Mm, mouse; Rn, rat; Hs, human; Pt, chimpanzee and Cf, dog.

G.P. Espino-Solis et al. / Veterinary Immunology and Immunopathology 132 (2009) 181–190 185

1) and 70% (peptide 2) compared with similar sequences ofthe CD11c aX I-domains of the corresponding homologoushuman, chimpanzee, mouse, rat and dog integrins. Amultiple alignment and the percentage of identity betweenthe peptides and the corresponding section of integrins ofother mammalian species are shown in Fig. 1D and E,respectively.

3.3. Screening of antibodies against CD11c synthetic peptides

in cell suspensions (FACS) and in horse lymph node in situ

(immunohistochemistry)

Cell suspensions of peripheral blood obtained from ahealthy horse were used to assess the antibody to peptide 2and the respective preimmune serum by means of flowcytometry (FACS profiles, Fig. 2A). Compared to the negativedot plots obtained with the corresponding preimmuneserum, antibodies to peptide 2 clearly provided a positivesignal. This is more clearly seen when overlapping thehistogram profiles obtained with each of these antibodies.Since we wanted to search positive cells within lymphoidtissues by means of immunohistochemistry, anti-peptidesera were used in conventional cryosections of horse lymph

node trying to identify cells in situ, and the results are shownin Fig. 2. The most important finding was that only the serumraised against peptide 2, the one corresponding to theMIDAS section of I-domain, identified positive cells withinthe tissue. Perhaps more relevant to the context of our workwas the finding that some of these cells exhibited a cleardendritic appearance even in situ (magnifications in Fig. 2D–F and G–I), despite the fact that these are conventionally cuttissue sections, not planar layers where the delicate dendriteprolongations can be more easily exposed. Indeed anti-peptide 2 (but not anti-peptide 1) antibodies recognizedcells in the lymph nodes whose morphology seems verysimilar to that normally seen in dendritic cells from otherspecies, from which it is concluded that this antibody doesrecognize an epitope of the integrin aX CD11c, as designed.Preinmune serum obtained from the same mice before theimmunization protocol was used as an isotype controlyielding negative results (Fig. 2B). In contrast, the antibodiesagainst peptide 1 were not able to clearly identify any cell inthe sections examined (Fig. 2C), from which it is concludedthat either these last antibodies do not recognize the foldedpeptide, or that this sequence in the 3D-structure of I-domain is not as exposed as the MIDAS domain might be.

Fig. 2. Flow cytometry of blood cell suspensions and immunohistochemistry of horse lymph node cryosections with antibodies obtained against CD11c

synthetic peptides. (A) Fluorescence activated cell sorter (FACS) of cell suspensions prepared from horse peripheral blood. Immunolabelling was done with

preimmune sera (upper left dot plots) as isotype control, and the antisera raised against peptide 2 (lower left dot plots). Overlapping histograms (far right-A)

illustrate the comparative staining obtained using the two antibodies. Immunohistochemistry of horse lymph node cryosections (B–I); (B) shows the results

with preimmune serum deemed as isotype control; (C) with anti-peptide 1 antibodies; and (D–I) the labelling with anti-peptide 2 antibodies. D and G are

enlarged in the next pictures (E, F and H, I) to illustrate cells with dendritic morphology. Arrowheads indicate cells with dendritic appearance in situ. Size

bars and magnifications are also indicated in the pictures.

G.P. Espino-Solis et al. / Veterinary Immunology and Immunopathology 132 (2009) 181–190186

The important point is that by using the strategy describedhere we can show, to the best of our knowledge for the firsttime, images of putative dendritic cells in situ within horselymph nodes, immunolabelled with antibodies generatedagainst segments of the I-domain of the CD11c integrin.

3.4. Staining of horse lymph node sections with PNA and anti-

peptide 2 serum

Trying to identify potential foci of B cells in horse lymphnodes, such as the germinal centres, serial tissue sectionswere stained with PNA or anti-peptide 2 serum. Labelling oflymph node frozen sections was carried out by theperoxydase technique using horseradish peroxidase-taggedPNA (HRP-PNA). As expected, PNA provided a strong andcharacteristic signal in the lymph node, likely depictinggerminal centres (Fig. 3C) whereas positive signal for CD11cwas obtained also inside the putative B cell follicles (Fig. 3Band D). In this regard, previous reports have also describedCD11c+ cells within follicles; first Liu’s group (Grouard et al.,

1996) discovered the so-called Germinal Centre DendriticCells (GCDCs), a bone marrow (BM) derived DC subsetdifferent from both tingible body macrophages (TBM), andfrom Follicular Dendritic Cells (FDCs, which are of stromallineage). More recently Nussenzweig’s group using CD11c-yellow-green fluorescent (EYFP) transgenic mice identifiedin situ DC only in the CD11c-EYFPhigh mice, describing threeDC populations associated to B cell follicles, one of theminside, stronglypositive for CD11c andverymotile (Lindquistet al., 2004). To better illustrate the histological location,serial tissue sections were stained with hematoxylin–eosin(HE, Fig. 3A), CD11c (Fig. 3B and D) and PNA (Fig. 3C),respectively.Again preimmuneserumwasused as anisotypecontrol antibody, yielding negative results (not shown).

3.5. Screening for crossreactivity of monoclonal antibodies to

human dendritic cells in horse lymphoid tissue

To the best of our knowledge there are no antibodies yetcapable of labelling specific markers on horse dendritic

Fig. 3. Horse lymph node cryosections stained with HE, PNA and anti-peptide 2 antibodies. Serial tissue sections of horse lymph node were stained with

hematoxylin–eosin (HE) (A) to provide a basic histological reference; with anti-peptide 2 antibody (B and D), and with PNA (C) to identify potential B cell

foci (putative germinal centres). Black bars illustrate size in microns and magnifications are indicated in each picture.

G.P. Espino-Solis et al. / Veterinary Immunology and Immunopathology 132 (2009) 181–190 187

cells. In the work of Mauel et al. (2006), the authors usedhorse cell suspensions and a panel of murine monoclonalantibodies against human dendritic cell markers (MHC-II,CD86, CD83 and mannose receptor) to show by flowcytometry that the horse cells they were growing in vitro,were dendritic cells. They showed crossreactivity of anti-human antibodies with horse dendritic cells. Similarly,Quesada-Pascual et al. (2008) used antibodies againsthuman DC markers (CD1a, CD86, MHC-II and langerin)trying to identify Langerhans cells in armadillo skin(Dasypus novemcinctus), showing crossreactivity only withanti-human CD86. We thus used a similar approach andfound that antibodies to human CD1a (Fig. 4A and B) andCD11c (Fig. 4C and D) did not label cells in horse lymphnode. Interestingly, the mAbs anti-CD86 and MHC-II useddid crossreact with horse tissue cells. In the case of anti-CD86, the staining was observed in cells surrounding theputative follicles (Fig. 4E and F). In contrast, MHC-IIstaining showed positive signal apparently within putativefollicles (Fig. 4G and H). The preimmune serum was used asnegative control (Fig. 4A and B), showing no staining, asexpected. In conclusion, here we also show that somemAbs that recognize human dendritic cells can crossreactwith cells in horse lymphoid tissues.

4. Discussion

This communication reports part of the results from amulti-stage project, aimed at the specific production ofhorse antibodies for possible biotechnological usage. Thefirst step of the project was the cloning and sequencing ofthe alpha-X subunit from CD11c/CD18 horse integrin(Espino-Solis et al., 2008). With the amino acid sequenceinformation obtained from cloning the horse CD11c it waspossible to obtain a three-dimensional model of the alpha-X I-domain using bioinformatics tools available in the web

servers (Fig. 1A of this communication). The three-dimensional model of the I-domain presented here iscongruent with the models of other known integrins. Infact, the folding of the protein follows the same scaffolddetermined for the other integrins (see Section 1)conserving the ‘‘conformational sign’’ mentioned byVorup-Jensen et al. (2003). Structural analysis of the I-domain facilitates the selection of two segments of theprimary structure for chemical synthesis. Peptides 1 and 2were synthesized, purified (Fig. 1B and C) and usedsuccessfully for the generation of mouse antibodies againstthese peptides. One of them (antibodies to peptide 2)showed first a positive FACS labelling of cell suspensionsobtained from peripheral horse blood, and also provedcapable of recognizing cells within the actual lymphoidtissue, some with prominent dendritic morphology even insitu. It is thus surmised that this epitope is accessible to theantibodies applied. In contrast, the serum obtained againstpeptide 1 did not label clearly any cell neither cells withdendritic appearance in horse lymph nodes. Perhaps this isso because the segment of alpha-helix corresponding tothis peptide is possibly hindered by other sections of theintegrin. It suffices to recall that our model (Fig. 1A) wasobtained using a segment of only 211 amino acids of aprotein that contains 1086 amino acids. This segment waschosen because it corresponds to the region of the integrinbelonging to the extra-cellular portion. However, thecontacts that this I-domain can establish with theremaining of the protein structure are not entirely known.It is thought that the mouse antibody obtained againstpeptide 1 does not bind to cells within the lymph nodebecause the epitope might not be accessible, although wehave no formal proof of this.

Although, the conventional, bone marrow (BM)-derivedDC but not the Follicular Dendritic Cells (FDCs, which are ofstromal origin) have been traditionally described in the T

Fig. 4. Screening for crossreactivity of antibodies to human dendritic cells in horse lymphoid tissue. Horse lymph node cryosections were used to assess

potential crossreactions of monoclonal antibodies against markers for human DCs. The following antibodies to human DC were used: CD1a (A and B), CD11c

(C and D), CD86 (E and F), MHC-II DR (G and H). Positive labelling is revealed by peroxidase with a chromogen yielding a dark-blue staining. An irrelevant

mouse monoclonal antibody was used as isotype control providing also negative staining like with CD1a antibody (not shown). Size bars and magnifications

are indicated in the pictures.

G.P. Espino-Solis et al. / Veterinary Immunology and Immunopathology 132 (2009) 181–190188

cell/Interdigitant Dendritic Cell (IDC) zone immediatelyaround the follicles, BM-derived DC have been alsoidentified recently in other anatomical locations. Forinstance, in a remarkable report, Grouard (Grouard et al.,

1996) discovered a novel type of CD11c+, BM derived DCinside B cell follicles called Germinal Centre DCs. This newDC subset was distinctively CD11c+ and was quite differentfrom both the tingible body macrophages (TBM, which are

G.P. Espino-Solis et al. / Veterinary Immunology and Immunopathology 132 (2009) 181–190 189

BM derived too) and from FDCs (which are not). However,from our preliminary results we cannot definitely claimthat the CD11c+ DCs we found in horse lymph nodefollicles, correspond exactly to the population described inhumans. More recently, Nussenzweig’s group used CD11c-yellow-green fluorescent (EYFP) mice and identified DC insitu only in the CD11c-EYFPhigh mice, reporting three DCpopulations associated to B cell follicles, one of them insidefollicles described as very bright for CD11c and as the mostmotile DC population (Lindquist et al., 2004).

Another interesting communication of this work is thepotential crossreactivity of mAbs directed to human cellsin horse lymphoid tissues in situ. Previous publicationsreported the use of anti-human mAbs in cell suspensionstrying to identify molecules on horse leukocytes andmononuclear cells from peripheral blood, by using flowcytometry (FACS) (Ibrahim and Steinbach, 2007). Some ofthese mAbs were shown to crossreact with horse cells.Mauel et al. (2006) prepared horse dendritic cells in vitro

and used mAbs against human MHC-II and CD86. TheirFACS analysis reported that cells were stained by thesetwo mAbs. Here, we have shown in situ by immunohis-tochemistry of horse lymph node sections, that the samemAbs to human CD86 and MHC-II apparently identifiedequine cells (Fig. 4). We also found that antibodies tohuman CD11c and CD1a did not recognize positive cellswithin this tissue (Fig. 4). These findings might not beunusual, since the sequence similarities of proteinsbetween the immune systems of horses and humansmight vary from 60 to 98% identity (Ibrahim andSteinbach, 2007). Thus, it is expected that some mAbsto human cells might not recognize exactly the sameepitopes in horse cells. There are reports of in situimmunohistochemistry in different horse tissues such asPeyer’s patches, lymph nodes, spleen, thymus and tonsilswith mAbs against human CD3, CD4, CD5 and CD8 of T andB cells (Kelley et al., 1997). However, to the best of ourknowledge none of these reports illustrate cells withtypical dendritic appearance. There are also other reportson the use of PNA identifying germinal centres in horselymphoid tissue (Merant et al., 2005).

We believe that the results reported here are somehownovel and interesting: first, they show that we can usemurine antibodies raised against peptides designed in themanner reported in this manuscript to assess in situputative dendritic cells directly within horse lymphoidtissues. We also show that some antibodies specific to cellsof the human immune system can crossreact in situ withthe corresponding target cells within horse tissues. Wehope that our findings will incite further characterizationof the horse immune system, a poorly studied animalmodel despite the fact that Emil von Behring startedalready a century ago the use of horse sera for humanserotherapy; which is still in use nowadays.

Acknowledgements

The authors are grateful to Dra. Maria Masri Daba andDr. Alejandro Rodrıguez Monterde from the VeterinarianMedical School of UNAM for providing us with the horsetissues. The scholarship 169946 of CONACyT (Consejo

Nacional de Ciencias y Tecnologia) given to GPES is alsorecognized. Partial economical support came also fromgrants of CONACyT and DGAPA-UNAM to LDP laboratory.

References

Alexandrov, N.N., Nussinov, R., Zimmer, R.M., 1996. Fast protein foldrecognition via sequence to structure alignment and contact capacitypotentials. Pac. Symp. Biocomput. 53–72.

Arnold, K., Bordoli, L., Kopp, J., Schwede, T., 2006. The SWISS-MODELworkspace: a web-based environment for protein structure homol-ogy modelling. Bioinformatics 22, 195–201.

Banchereau, J., Briere, F., Caux, C., Davoust, J., Lebecque, S., Liu, Y.J.,Pulendran, B., Palucka, K., 2000. Immunobiology of dendritic cells.Annu. Rev. Immunol. 18, 767–811.

Banchereau, J., Steinman, R.M., 1998. Dendritic cells and the control ofimmunity. Nature 392, 245–252.

Bilsland, C.A., Diamond, M.S., Springer, T.A., 1994. The leukocyte integrinp150,95 (CD11c/CD18) as a receptor for iC3b. Activation by a hetero-logous beta subunit and localization of a ligand recognition site to theI domain. J. Immunol. 152, 4582–4589.

Calderon-Aranda, E.S., Olamendi-Portugal, T., Possani, L.D., 1995. The useof synthetic peptides can be a misleading approach to generatevaccines against scorpion toxins. Vaccine 13, 1198–1206.

Calderon-Aranda, E.S., Selisko, B., York, E.J., Gurrola, G.B., Stewart, J.M.,Possani, L.D., 1999. Mapping of an epitope recognized by a neutraliz-ing monoclonal antibody specific to toxin Cn2 from the scorpionCentruroides noxius, using discontinuous synthetic peptides. Eur. J.Biochem. 264, 746–755.

Espino-Solis, G.P., Osuna-Quintero, J., Possani, L.D., 2008. Molecularcloning and characterization of the alphaX subunit from CD11c/CD18 horse integrin. Vet. Immunol. Immunopathol. 122, 326–334.

Flores-Langarica, A., Meza-Perez, S., Calderon-Amador, J., Estrada-Gar-cia, T., Macpherson, G., Lebecque, S., Saeland, S., Steinman, R.M.,Flores-Romo, L., 2005. Network of dendritic cells within the mus-cular layer of the mouse intestine. Proc. Natl. Acad. Sci. U.S.A. 102,19039–19044.

Flores-Romo, L., 2001. In vivo maturation and migration of dendritic cells.Immunology 102, 255–262.

Grouard, G., Durand, I., Filgueira, L., Banchereau, J., Liu, Y.J., 1996. Den-dritic cells capable of stimulating T cells in germinal centres. Nature384, 364–367.

Hynes, R.O., 2002. Integrins: bidirectional, allosteric signaling machines.Cell 110, 673–687.

Ibrahim, S., Steinbach, F., 2007. Non-HLDA8 animal homologue sectionanti-leukocyte mAbs tested for reactivity with equine leukocytes. Vet.Immunol. Immunopathol. 119, 81–91.

Jimenez-Flores, R., Mendez-Cruz, R., Ojeda-Ortiz, J., Munoz-Molina, R.,Balderas-Carrillo, O., de la Luz Diaz-Soberanes, M., Lebecque, S., Sae-land, S., Daneri-Navarro, A., Garcia-Carranca, A., Ullrich, S.E., Flores-Romo, L., 2006. High-risk human papilloma virus infection decreasesthe frequency of dendritic Langerhans’ cells in the human femalegenital tract. Immunology 117, 220–228.

Kelley, L.C., Mahaffey, E.A., Bounous, D.I., Antczak, D.F., Brooks Jr., R.L.,1997. Detection of equine and bovine T- and B-lymphocytes informalin-fixed paraffin-embedded tissues. Vet. Immunol. Immuno-pathol. 57, 187–200.

Lindquist, R.L., Shakhar, G., Dudziak, D., Wardemann, H., Eisenreich, T.,Dustin, M.L., Nussenzweig, M.C., 2004. Visualizing dendritic cell net-works in vivo. Nat. Immunol. 5, 1243–1250.

Luo, B.H., Carman, C.V., Springer, T.A., 2007. Structural basis of integrinregulation and signaling. Annu. Rev. Immunol. 25, 619–647.

Mauel, S., Steinbach, F., Ludwig, H., 2006. Monocyte-derived dendriticcells from horses differ from dendritic cells of humans and mice.Immunology 117, 463–473.

Mellman, I., 2007. Private lives: reflections and challenges in under-standing the cell biology of the immune system. Science 317, 625–627.

Merant, C., Messouak, A., Cadore, J.L., Monier, J.C., 2005. PNA-bindingglycans are expressed at high levels on horse mature and immature Tlymphocytes and a subpopulation of B lymphocytes. Glycoconj. J. 22,27–34.

Pulendran, B., Banchereau, J., Maraskovsky, E., Maliszewski, C., 2001.Modulating the immune response with dendritic cells and theirgrowth factors. Trends Immunol. 22, 41–47.

Quesada-Pascual, F., Jimenez-Flores, R., Flores-Langarica, A., Silva-San-chez, A., Calderon-Amador, J., Mendez-Cruz, R., Limon-Flores, A.Y.,

G.P. Espino-Solis et al. / Veterinary Immunology and Immunopathology 132 (2009) 181–190190

Estrada-Parra, S., Santos-Argumedo, L., Estrada-Garcia, I., Flores-Romo, L., 2008. Characterization of langerhans cells in epidermalsheets along the body of Armadillo (Dasypus novemcinctus). Vet.Immunol. Immunopathol. 124, 220–229.

Rost, B., Yachdav, G., Liu, J., 2004. The PredictProtein server. Nucleic AcidsRes. 32, W321–326.

Sheppard, E.A.R.C., 1989. Solid Phase Peptide Synthesis: A PracticalApproach. IRL Press, Oxford, New York and Tokyo.

Steinbach, F., Deeg, C., Mauel, S., Wagner, B., 2002a. Equine immunology:offspring of the serum horse. Trends Immunol. 23, 223–225.

Steinbach, F., Mauel, S., Beier, I., 2002b. Recombinant equine interferons:expression cloning and biological activity. Vet. Immunol. Immuno-pathol. 84, 83–95.

Ueno, H., Klechevsky, E., Morita, R., Aspord, C., Cao, T., Matsui, T., DiPucchio, T., Connolly, J., Fay, J.W., Pascual, V., Palucka, A.K., Bancher-eau, J., 2007. Dendritic cell subsets in health and disease. Immunol.Rev. 219, 118–142.

Vorup-Jensen, T., Ostermeier, C., Shimaoka, M., Hommel, U., Springer, T.A.,2003. Structure and allosteric regulation of the alpha X beta 2 integrinI domain. Proc. Natl. Acad. Sci. U.S.A. 100, 1873–1878.