Krüppel-Like Factors Integrin Targeted by 7Β Leukocyte

Home , FHL3, ID3 (gene)

Leukocyte β7 Integrin Targeted by Krüppel-like Factors Melanie Alles, Gleb Turchinovich, Pumin Zhang, Wolfgang Schuh, Fabien Agenès and Jörg Kirberg This information is current as of September 29, 2021. J Immunol 2014; 193:1737-1746; Prepublished online 11 July 2014; doi: 10.4049/jimmunol.1302613 http://www.jimmunol.org/content/193/4/1737 Downloaded from

Supplementary http://www.jimmunol.org/content/suppl/2014/07/11/jimmunol.130261 Material 3.DCSupplemental

References This article cites 69 articles, 32 of which you can access for free at: http://www.jimmunol.org/ http://www.jimmunol.org/content/193/4/1737.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists by guest on September 29, 2021

• Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2014 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Leukocyte b7 Integrin Targeted by Kruppel-like€ Factors

Melanie Alles,* Gleb Turchinovich,†,‡ Pumin Zhang,x Wolfgang Schuh,{ Fabien Agene`s,‖,#,1 and Jo¨rg Kirberg*

Constitutive expression of Kruppel-like€ factor 3 (KLF3, BKLF) increases marginal zone (MZ) B cell numbers, a phenotype shared with mice lacking KLF2. Ablation of KLF3, known to interact with serum response factor (SRF), or SRF itself, results in fewer MZ B cells. It is unknown how these functional equivalences result. In this study, it is shown that KLF3 acts as transcriptional repressor for the leukocyte-specific integrin b7 (Itgb7, Ly69) by binding to the b7 promoter, as revealed by chromatin immunoprecipitation. KLF2 overexpression antagonizes this repression and also binds the b7 promoter, indicating that these factors may compete for target sequence(s). Whereas b7 is identified as direct KLF target, its repression by KLF3 is not connected to the MZ B cell increase because b7-deficient mice have a normal complement of these and the KLF3-driven increase still occurs when b7 is deleted. Despite this, KLF3 overexpression abolishes lymphocyte homing to Peyer’s patches, much like b7 deficiency does. Furthermore, KLF3 expression alone overcomes the MZ B cell deficiency when SRF is absent. SRF is also dispensable for the Downloaded from

KLF3-mediated repression of b7. Thus, despite the shared phenotype of KLF3 and SRF-deﬁcient mice, cooperation of these factors appears neither relevant for the formation of MZ B cells nor for the regulation of b7. Finally, a potent negative regulatory feedback loop limiting KLF3 expression is shown in this study, mediated by KLF3 directly repressing its own gene promoter. In summary, KLFs use regulatory circuits to steer lymphocyte maturation and homing and directly control leukocyte integrin expression. The Journal of Immunology, 2014, 193: 1737–1746. http://www.jimmunol.org/

ruppel-like€ factors (KLFs) are zinc finger family tran- KLF3 may act exclusively in concert with other transcriptional scription factors with 17 members described for mam- regulators. K malian cells (1). KLFs have a highly conserved DNA In lymphocytes, KLFs play important roles in their development binding domain comprised of three C2H2 zinc fingers near or at the and function (7, 9, 13–20). When constitutively expressed under C terminus and share similarity to the transcription factors of the the B cell–specific CD19 promoter, KLF3 increases maturation specificity protein family (2–5). The KLF zinc fingers are thought toward marginal zone (MZ) B cells, resulting in a 5- to 10-fold to mediate binding to CACCC elements and GC-rich DNA increase of this subset, whereas, accordingly, MZ B cells were sequences in target genes. reduced in KLF3-deficient mice (7, 9). Interestingly, absence of by guest on September 29, 2021 In Drosophila Schneider cells, KLF3 (BKLF) had been shown KLF2 in B cells results in an increase of MZ B cells, mimicking to act as a weak transcriptional activator (6). However, more re- the effect of constitutive KLF3 expression (15, 16, 19). Although cent data in murine cells implicate KLF3 as a repressor of tran- the actual target genes mediating this effect are unknown, it has scription (7–9). Indeed, KLF3 has been shown to interact with the been demonstrated that the lymphocyte phenotypes of KLF2- common transcriptional corepressors C-terminal binding protein 2 deficient mice result from effects on the positioning, recircula- (CtBP2) and four and a half LIM domain protein 3 (FHL3) (8, 10, tion, and/or tissue-homing ability of these cells (13, 14, 19, 21– 11). A further KLF3-interacting transcription factor is serum 24). However, not all outcomes of KLF2 deficiency resulted response factor (SRF), as identified in muscle cells (12). Thus, from cell-autonomous processes (18). Integrins perform crucial functions for the positioning of leukocytes by acting both as signaling and signaling-responsive ad- *Division of Immunology (3/3), Paul-Ehrlich-Institut, 63225 Langen, Germany; hesion molecules (25). The functions of the leukocyte-specific b2 †Department of Biomedicine, Laboratory of Developmental Immunology, 4058 Basel, x b Switzerland; ‡Basel University Children’s Hospital, 4031 Basel, Switzerland; Department (CD18) and 7 (Ly69) integrin subunits have been elucidated of Molecular Physiology and Biophysics, Baylor College of Medicine, Houston, TX initially by blocking studies and subsequently by the respective { 77030; Division of Molecular Immunology, Department of Internal Medicine III, gene-deficient mice. This led to a mouse model of leukocyte ad- Nikolaus-Fiebiger-Center, University of Erlangen-Nurnberg,€ 91054 Erlangen, Germany; ‖INSERM U743, Montreal, Quebec H2X 1P1, Canada; and #INSERM ADR Paris V Saint hesion deficiency (LFA-A immunodeficiency, OMIM 116920) and Anne, 75014 Paris, France specific defects in the migration of leukocytes to tissues such as 1Current address: Office for Science and Technology, Consulate General of France, skin and gut, respectively (26–31). Indeed, whereas leukocyte Los Angeles, CA. migration via the afferent lymph and within lymph nodes is not Received for publication September 30, 2013. Accepted for publication June 6, 2014. affected in the absence of all integrins, their extravasation is Address correspondence and reprint requests to Dr. Jo¨rg Kirberg, Paul-Ehrlich- crucially dependent on them (32). With respect to b7 integrin, this Institut, Division of Immunology (3/3), D-63225 Langen, Germany. E-mail address: subunit can either pair with a (CD49d) to form lymphocyte [email protected] 4 Peyer’s patch (PP) adhesion molecule-1 (a /b ), mediating bind- The online version of this article contains supplemental material. 4 7 ing to VCAM-1, mucosal addressin cell adhesion molecule 1 Abbreviations used in this article: ChIP, chromatin immunoprecipitation; CtBP2, a C-terminal binding protein 2; KLF, Kruppel-like€ factor; MAdCAM-1, mucosal (MAdCAM-1; addressin), and fibronectin, or pair with E (CD103), addressin cell adhesion molecule 1; MLn, mesenteric lymph node; MZ, marginal a heterodimer found particularly among intestinal intraepithelial zone; PP, Peyer’s patch; QRT-PCR, quantitative real-time PCR; RIPA, radioimmu- lymphocytes mediating binding to E-cadherin. Thus, b expres- noprecipitation assay; SRF, serum response factor. 7 sion is considered a hallmark of mucosal lymphocytes because Copyright Ó 2014 by The American Association of Immunologists, Inc. 0022-1767/14/$16.00 a4/b7 allows for binding to the high endothelial venules of gut www.jimmunol.org/cgi/doi/10.4049/jimmunol.1302613 1738 KLF2/3-REGULATING LEUKOCYTE b7 INTEGRIN

PP and mesenteric lymph nodes (MLn), and endothelial cells of streptomycin (Life Technologies), b-ME (50 mM), Gln (4 mM), 0.03% the lamina propria, all specifically expressing MAdCAM-1 (33), (w/v) Primatone RL (Quest, 5x59051, 10-kDa ultrafiltrate) (52), 5 mg/ml insulin (Sigma-Aldrich I5500), and 2% FCS (unless indicated differently) and aE/b7 mediating the retention within the tissue by binding to at 5–7.5% CO2, 37˚C. E-cadherin, expressed in the gut epithelial layer. Whereas aE (CD103) only pairs with b7, a4 (CD49d) can pair with either b7 or Retroviral transduction b a b 1 (CD29), the latter termed VLA-4 ( 4/ 1, CD49d/CD29), medi- Retroviral vector particles were obtained by transient transfection of ret- ating binding to VCAM-1, MAdCAM-1, and fibronectin. Whereas roviral vector plasmid DNA into Plat-E producer cells (53). Briefly, the day 3 6 binding specificities thus overlap, the ratio of a4/b7 and a4/b1 de- prior to transfection, 5 10 Plat-E cells were plated in a T75 flask in 8– 10 ml medium (10% FCS). The next afternoon, chloroquine (25 mM final) termines lymphocyte-homing specificity because a4/b7 preferentially (54) was added, and 1 h later cells were transfected using calcium- binds to MAdCAM-1 and a4/b1 has higher affinity for VCAM-1 phosphate [to 1 ml 20 mg plasmid DNA in 0.295 M CaCl2, 1 ml 50 mM (34–36). The relative expression of these integrins is set transcrip- N,N-bis(2-hydroxyethyl)-2-aminoethanesulfonic acid (BES), 280 mM tionally but also depends on intracellular pairing preferences. Ap- NaCl, 1.5 mM Na2HPO4 (pH 7.16) was added, gassed with air for 20 s, and parently, under nonstimulated conditions, the expression of a4 and, immediately the mixture was added to the cells] (55). The next morning, medium was exchanged. Virus particle-containing supernatants were taken particularly, b1 is limited, such that the preferentially associating a4/b1 a b 42, 50, 66, and 72 h after transfection. They were made cell free (10 min, heterodimer does not fully outcompete the formation of 4/ 7 (37). 3300 rpm, 4˚C), snap frozen in liquid nitrogen, and stored at 280˚C. For Given the discrete binding preferences of the various integrin transduction of some cells, retroviral particles were pseudotyped with the heterodimers, their relative expression has pronounced effects on vesicular stomatitis virus glycoprotein G by adding plasmid pVPack-VSV-G the tissue specificity of extravasation. Consequently, there has been (Stratagene) to the transfection. For transduction, 1 3 105 target cells/well were seeded in 12-well interest in the characterization of the promoter elements of various Downloaded from plates. Virus-containing supernatants were thawed on ice and used as such integrins and the molecules regulating expression. Factors of the or aliquoted (1.5 ml), centrifuged (60 min, 14,000 rpm, 4˚C), most su- specificity protein family of transcription factors have been directly pernatant aspirated, and the remaining liquid was resuspended and pooled shown to bind to several integrin promoters (38–42). Notably, the to ∼1 ml vol. Then polybrene was added to 5 mg/ml. Medium was removed promoter for b has been identified, but there has been a paucity of from target cells, and the virus-containing solution was added, performing 7 a spin infection (60 min, 1,800 rpm, 37˚C). Thereafter, cells were incu- defined factors that control expression of this lymphocyte-specific bated normally, changing medium 1 h later. Cells were split as needed, and integrin (43, 44). on day 3 after transduction they were analyzed by FACS. http://www.jimmunol.org/ In this study, it is shown that KLF3 and KLF2 both directly tar- In vivo lymphocyte-homing assays get the promoter of b7 integrin, thereby contributing to the regulation of this gene in lymphocytes. Their relative expression levels de- Donor lymphocytes were obtained from spleen (erythrocytes lysed using fine whether repression or activation occurs. Furthermore, we show Tris-buffered ammonium chloride) and peripheral (all non-i.p.) lymph that KLF3 mediates a negative regulatory feedback loop directly nodes of the indicated strains by passing through nylon meshes into PBS b with 0.5% BSA (EQBAC62; Europa Bioproducts). Mixtures (1:1) based on on its own promoter. However, whereas KLF3 represses 7 and cell counts were prepared and washed into PBS, and 107–3 3 107 cells were thereby abolishes lymphocyte homing to PP and MLn, this is not adoptively transferred into B6.Ly-5a recipient mice i.v. Recipients were the mechanism behind the increased MZ B cell compartment in analyzed 15 h later, collecting PP from equivalent regions of the intestine, CD19:KLF3 transgenic mice, as this phenotype is shown to be peripheral lymph nodes, MLn, and spleen. Blood lymphocytes were pu- by guest on September 29, 2021 rified by Ficoll density gradient centrifugation. b7 independent. Likewise, KLF3 overcomes the paucity of MZ B cells in mice deficient for SRF, thus showing that MZ B cells FACS analyses are not depending on the combined action of these two factors. Cells were isolated from the indicated organs, as described above. For samples from spleen or bone marrow, erythrocytes were lysed (Tris-buffered Materials and Methods ammonium chloride). Adherent cells from tissue culture were detached Mice with trypsin/EDTA. Staining was performed in PBS with 0.5% BSA using reagents at predetermined optimal dilution, maximizing signal to noise, and CD19:KLF3 transgenic mice [C57BL/6J-TgN(CD19:KLF3)2Bii] were adding 2.4G2 (CD16/32) hybridoma supernatant or normal mouse Ig to tm1(cre/ERT2)Pzg described previously (7). The B6;129P-Klf3 /J (Klf3-GCE, limit unspecific staining. Dead cells were excluded by staining with Sytox- derived within the GUDMAP project [http://www.gudmap.org]) (45, 46), blue (Invitrogen). Abs were from BD Biosciences and eBioscience, or tm1Ddg tm1(cre)Cgn B6.129-Srf /J (47), B6.129P2(C)-Cd19 /J (48, 49), B6.SJL- prepared by protein G affinity chromatography, followed by derivatization a b tm1Cgn Ptprc Pepc /BoyJ, C57BL/6-Itgb7 /J (30), and C57BL6/J (B6) mouse using standard techniques. Flow cytometric analyses were performed on strains were obtained from The Jackson Laboratory (Bar Harbor, ME). a LSRII-SORP (BD Biosciences) instrument with Diva 4.1-6.1 software. a/b a b B6.Ly-5 mice derive from intercrosses of B6 and B6.SJL-Ptprc Pepc /BoyJ Data were analyzed using FlowJo 7.6.3-7.6.5 software. Unless specified a b a mice. Recipient B6.SJL-Ptprc Pepc /BoyCrl (B6.Ly-5 ) mice were from differently, numbers indicate the proportion of cells within the given gate or Charles River Italy. Animal work and breeding were performed at the region. The p values (Student t test) from pairwise comparisons are indi- mouse facilities in Montreál and Langen. PCR protocols/primer sequences cated within figures or their legends. for genotyping are available upon request. Studies on animals were performed concurrent to the local legal requirements. Chromatin immunoprecipitation Cloning Cells (2 3 107) were washed into PBS and fixed with formaldehyde (0.75%, 15 min). Fixation was stopped with glycine (125 mM final). Fixed Tagged versions of KLF3 were generated by PCR using vector pMT2/BKLF cells were washed into PBS, then lysed (1 h, 4˚C, in 1.2 ml lysis buffer (6) or derivatives as template (see Supplemental Fig. 1A). In the KLF3- containing 50 mM HEPES-KOH [pH 7.5], 140 mM NaCl, 1 mM EDTA DZF constructs, the zinc finger domain of KLF3 is deleted, such that 271 [pH 8.0], 1% Triton X-100, 0.1% Na-deoxycholate, 1% SDS, protease a TAA stop codon follows amino acid Thr of KLF3 (AAA93256.1). In inhibitors [cOmplete ULTRA; Roche]). After sonication to an average D the KLF3- DL mutant, the PVDLT motif in KLF3 was inactivated by fragment size of 500–1000 bp (Bandelin Sonopuls HD2200, Sonotrode 63 . 64 . Asp - Ala and Leu - Ser, as described (8). All expression constructs MS73, 10-s pulses, 36–40% power), the lysates were cleared (7,640 3 g, were verified by sequencing. They were shuttled into pMYc-IRES-GFP 5 min, 4˚C, then 18,500 3 g, 2 min), aliquoted, and snap frozen. (50) or a derived pMYc-IRES-tdTomato retroviral vector backbone for For chromatin immunoprecipitation (ChIP), protein G magnetic beads expression in target cells by retroviral transduction. pBMN-LZRS-IRES- (NEB) were blocked in radioimmunoprecipitation assay buffer (RIPA) GFP (51) served as vector backbone for the KLF2 construct. containing 0.1 mg/ml BSA and one-tenth volume Roti-Block (Carl Roth). 6 Tissue culture Lysates representing the equivalent of 1.33 3 10 cells were diluted 1/5 with RIPA buffer, and then either goat anti-KLF3 (Acris) or rabbit anti- Cell culture was in IMDM (Biochrom, purchased as powder) supplemented KLF2 (56) was added (1–10 mg per 400 ml diluted lysate). Then, 15-ml with MEM nonessential amino acids, 100 IU/ml penicillin, 100 mg/ml beads were added, agitating overnight at 4˚C. Thereafter, beads were The Journal of Immunology 1739 washed twice each at low (0.05% SDS, 1% Triton X-100, 2 mM EDTA [pH 8.0], 150 mM NaCl, 20 mM Tris [pH 8.0]) and high salt concentration (0.05% SDS, 1% Triton X-100, 2 mM EDTA [pH 8.0], 500 mM NaCl, 20 mM Tris [pH 8.0]). DNA was eluted from beads by adding 120 ml1% SDS, 100 mM NaHCO3 (15 min, 30˚C). The eluate was diluted with RIPA buffer (to 400 ml), adding RNase, then proteinase K, to liberate the DNA and to reverse the formaldehyde fixation (65˚C, 4–5 h). Finally, the DNA was purified by phenol/chloroform extraction. Glycogen was added as carrier for ethanol precipitation. Recovered DNA was resuspended in water and stored at 4˚C until analysis. To this end, quantitative real-time PCR (QRT-PCR) was performed on a Light Cycler 480 instrument (Roche) and the MESA Blue qPCR MasterMix Plus for SYBR Assay no ROX (Euro- gentec), using the absolute quantification method, to determine the relative DNA abundance for the primer-defined DNA region in the given ChIP sample. Three primer pairs for the b7 integrin promoter, one pair for the KLF3 1b promoter region (57) and a corresponding control (210 kb to the KLF3 1a promoter), and one general control primer pair amplifying an irrelevant mouse genomic region (Furin gene) were used. Primer sequences are as follows: 380_210 kb_KLF3_s, 59-CGG GCA ATG AGG CAT CAG GGT-39; 381_210 kb_KLF3_a, 59-GGC TCA CTT GGG CTC ACC ATC T-39; 382_Itgb7_site3_s, 59-CGG GCT GCT CTA GCT GCA CTC-39; 383_Itgb7_site3_a, 59-CAG CCC GAG GTG GAG GCT TT-39; 384_Itgb7_- site1_s, 59-GCA GCC CCT TGC TAC CTG TGT G-39; 385_Itgb7_site1_a, 59- Downloaded from TCA CCA GAA CAG CGC CGG AT-39; 406_Itgb7_site2_s, 59-GAG CAG GAT GGG GTG TTC AA-39; 407_Itgb7_site2_a, 59-ACC AGC AGT GTA CAA GAG CC-39; 414_Bklf_1b_s, 59-CTG GGT GTG GGC AGA ATC TT- 39; 415_Bklf_1b_a, 59-CTC CCC TCT CCG TGA ACA AG-39; 424_control4 (Furin)_s, 59-GCT CCT GTA GTC CTA AGC GG-39; 425_control4(Furin)_a, 59-TGG ATG TGA TTG AGG CTG GG-39. The suitability of primer pairs for QRT-PCR was tested by amplification of titrated amounts of genomic mouse DNA. All primer pairs had com- http://www.jimmunol.org/ parable amplification efficiency (,62 cycle thresholds) and gave a consistent product melting curve. Values reported represent the amount of recovered DNA relative to the input DNA, as present in lysates before FIGURE 1. KLF3 represses b7 integrin expression in vivo. (A) Splenic ChIP, performing QRT-PCR determinations in duplicate. B cells of B6 and CD19:KLF3 transgenic littermates were analyzed for

IgM, CD21/35, and b7 integrin expression. Shown are B cells as gated on Results basis of being CD19+CD932. Differentiation into follicular and MZ B KLF3 represses b7 integrin expression in vivo cells is based on the gates shown on the left. Gray shades have been positioned into histograms at the same marker expression level to aid in the Comparing follicular and MZ B cells of normal and CD19:KLF3 b visualization of changes. Data are representative for eight B6 and CD19: by guest on September 29, 2021 transgenic mice by mRNA array analysis, 7 integrin was strongly KLF3 transgenic mice, each. (B) Shown are geometric means and SDs for downregulated in KLF3 transgenic cells while being somewhat b7 expression among follicular and MZ B cells from B6 and CD19:KLF3 higher expressed in the absence of KLF3 (7, 9). Thus, it was tested transgenic littermates, as specified in (A). Because data derive from several whether KLF3 represses b7 as determined by cell surface staining. independent experiments using different instrument settings and fluo- b b Indeed, downregulation of b7 is observed among splenic follicular rochromes, 7 expression data were first normalized on basis of 7 ex- B cells of CD19:KLF3 transgenic mice. Notably though, this is the pression on follicular B cells. The multiplicator for this was derived using b case for many, but clearly not all, follicular B cells (Fig. 1A, left the mean value of 7 expression on follicular cells from all B6 animals within a given staining/experiment. The p values are indicated within the histogram panels). With respect to MZ B cells, expressing less b7 normally, all cells appear to be affected (Fig. 1A, right histogram figure. panels). KLF3 can act despite its PVDLT motif being inactivated, lacking KLF3 represses and KLF2 upregulates b integrin expression 7 the potential to recruit CtBP2 (8). in P815 mastocytoma cells KLF2 partially appears to possess functions that are opposing to b To further characterize the ability of KLF3 to regulate 7 ex- those of KLF3, such as in the generation of MZ B cells or lower b7 pression, we resorted to a murine mastocytoma cell line, P815 expression in KLF2-deficient B cells (15, 16, 19). Indeed, upon b (58), known to express 7 integrin. Consistent with the decrease KLF2 overexpression in P815 cells, b7 expression is increased b in 7 expression in B cells of CD19:KLF3 transgenic mice, ex- (Fig. 2B). The opposing functions of KLF3 and KLF2 are thus b pression of KLF3 in P815 cells strongly decreases 7 expression reproducible in this cell line. (Fig. 2A; for an overview of the various constructs, see Supplemental b Fig. 1A). In contrast, expressing a truncated version of KLF3 KLF2 and KLF3 work antagonistically on 7 integrin expression lacking its zinc fingers, KLF3-DZF, does not change b7 expression. This demonstrates that KLF3 requires its zinc finger domain for To determine how KLF3 and KLF2 interact, they were simulta- repression, apparently utilizing this domain for binding to pro- neously transduced into P815 cells, followed by monitoring b7 moter elements in cellular DNA. Additional data confirmed that integrin expression. As above, KLF3-transduced cells exhibit re- KLF3-DZF localized to the nucleus as does the full-length protein duced, whereas KLF2-transduced cells exhibit elevated b7 integrin (Supplemental Fig. 1B). Because KLF3 had been demonstrated to levels (Fig. 3, top center histogram panels). When using a three- interact with CtBP2 via a defined motif (8, 11), it was possible to dimensional plot, it appears that the relative level of the two test whether this interaction is also relevant for the repression of factors, as determined by the respective marker gene expression b7. Interestingly, expression of the CtBP interaction domain mu- GFP and tdTomato, determines the level of b7 expression (data not tant of KLF3, KLF3-DDL, still leads to significant downregulation shown). Because b7 integrin levels of KLF2 and KLF3 double- of b7 expression in P815 cells (Fig. 2A). Thus, in this setting, expressing cells, as gated in Fig. 3 (top right panel), approached 1740 KLF2/3-REGULATING LEUKOCYTE b7 INTEGRIN

known to express high levels of b7 and lacking b1 integrin expression (59–62). Downregulation of b7 integrin expression is evident upon KLF3 expression in these cells, whereas the zinc ﬁnger domain-deﬁcient variant KLF3-DZF has no effect (Fig. 4, upper center panels). An attempt was also made to increase b7 expression by introduction of KLF2; however, TK1 cells express relatively high amounts of KLF2 endogenously, and no increase in b7 expression is observed when introducing the KLF2-encoding retrovirus (Fig. 4, right panel). As TK1 cells lack expression of b1, b7 repression is accompanied by a decrease in a4 integrin surface expression (Fig. 4, lower panels).

KLF2 and KLF3 directly bind to the b7 integrin promoter As demonstrated above, KLF3 and KLF2 act antagonistically in the regulation of b7 integrin expression. In the case of TK1, lack of b7

FIGURE 2. KLFs affect b7 integrin expression in P815 mastocytoma expression on the cell surface could have also been caused by the cells. (A) P815 cells expressing KLF3 or variously altered versions of it, as repression of a4 expression. The results in P815, remaining a4 schematized above each dataset, were obtained by retroviral transduction. positive following introduction of KLF3 while b7 expression van- b They were analyzed for 7 integrin cell surface expression. Transduced ishes, would argue against this. In any case, to determine whether Downloaded from and nontransduced cells are those being marker gene positive (GFP+)or b 2 KLF3 was directly involved in the regulation of 7 integrin ex- being GFP , respectively, and stem from the same culture. Transduction pression, ChIP experiments were performed. As TK1 cells express B A by empty vector served as additional control. ( ) Similar to ( ), however, only low levels of endogenous KLF3 (data not shown), they were cells were transduced with a retrovirus encoding KLF2. (A and B) The cell also transduced to express higher levels. Using a KLF3-speciﬁc surface expression of a4 integrin was monitored at the same time, but no alterations could be detected (data not shown). Each experimental condi- antiserum in ChIP, the precipitated material is enriched for DNA b tion was reproduced independently at least two times, giving comparable corresponding to two nonoverlapping regions of the 7 promoter http://www.jimmunol.org/ results. region (∼2270 to 2130 nt and +920 to +1050 nt relative to the transcriptional start site of b7) (44) (Fig. 5A). This binding of KLF3 those of nontransduced or empty vector-transduced control cells to the b7 promoter is detectable both in normal TK1 and, yielding (Fig. 3, lower panels), this indicates that with the KLF3 expres- increased signals, in TK1 cells overexpressing KLF3. Because KLF3 sion level attained, KLF3 exerted a dominant effect. Notably, this and KLF2 act antagonistically, as demonstrated above, the ChIP occurs even though the proteins are expressed from heterologous was repeated using a KLF2-speciﬁc antiserum. As TK1 cells express promoters, excluding the possibility of acting by regulating each KLF2 endogenously, there was no need for overexpression. Indeed, other’s expression. KLF2 also binds to the b7 promoter regions (Fig. 5B). Together,

these results demonstrate that KLF3 and KLF2 both directly bind to by guest on September 29, 2021 KLF3 represses b integrin expression in TK1 T cell lymphoma 7 the b integrin promoter, giving a molecular explanation for their cells 7 antagonistic function in the regulation of b7 expression. The data above show that KLF3 and KLF2 work antagonistically KLF3 expression impacts lymphocyte homing in the regulation of b7 integrin expression in P815 mastocytoma cells. To analyze whether these functions are also relevant for As KLF3 directly represses b7 expression, we tested whether this lymphoid cells, the murine T lymphoma cell line TK1 was used, would affect lymphocyte homing. To this end, medium-term (15-h)

FIGURE 3. b7 integrin expression following KLF3 and KLF2 coexpression in P815 mastocytoma cells. P815 cells were transduced simultaneously with KLF3 and KLF2 encoding vectors (top panels) or the corresponding empty vector controls (bottom panels). GFP served as marker for KLF2 and tdTomato + + as marker for KLF3 expression. b7 integrin expression was monitored among nontransduced (neg.), single KLF3 (tdTomato ), single KLF2 (GFP ), and both KLF3- and KLF2-expressing (tdTomato+GFP+) cells following transduction. Numbers within dot plots indicate the proportion of transduced cells for the given retroviral construct(s), as determined by marker gene expression, with the gating shown. The upper numbers in histograms report the geometrical mean of b7 expression of the population as shown, and the lower numbers report the corresponding outcome of a similar independent second experiment; gray shades have been positioned at the same b7 expression level to aid in the visualization of changes. The Journal of Immunology 1741

FIGURE 4. KLF3 expression affects b7 and a4 integrin expression in TK1 T lymphoma cells. TK1 cells expressing KLF3 or a variant of KLF3 lacking its zinc ﬁnger domain, as schematized above the dataset, or expressing KLF2, were obtained by retroviral transduction. They were analyzed for b7 (top row) and a4 integrin (bottom row) cell surface expression. Transduced and nontransduced cells are those being marker gene positive (GFP+) or being GFP2, respectively, and stem from the same cultures. Transduction by empty vector served as additional control. The experiment was performed twice, giving comparable results.

migration assays were performed because our focus was on B cells fined by staining for various marker combinations (CD1d versus (29). Equivalent numbers of lymphocytes obtained from peripheral CD23, CD21/35 versus IgM, or CD21/35 versus CD23; Fig. 7A, lymph nodes and spleens of B6.Ly-5a/b control and various Ly-5b 7C, and data not shown). This could also be confirmed by 2/2 experimental mice, either B6, CD19:KLF3, or Itgb7 , were immunohistological analysis (Supplemental Fig. 3). Moreover, Downloaded from a transferred into B6.Ly-5 recipients. As expected for cells having when the CD19:KLF3 transgene is bred onto the b7-deficient reduced or lacking b7 expression (29–31), homing to PP and MLn background, the presence of the transgene still leads to an en- was severely abolished (Fig. 6, Supplemental Fig. 2). For cells richment for MZ B cells, demonstrating that the effects of KLF3 from CD19:KLF3 transgenic mice, this only affected B cells, can be clearly separated with respect to the regulation of b7 versus consistent with the specificity of transgene expression. Interest- the induction of MZ B cell maturation (Fig. 7B, 7C).

ingly, the remaining fraction of CD19:KLF3 transgenic cells http://www.jimmunol.org/ SRF is irrelevant both for KLF3 repressing b and driving MZ detected in PP and MLn showed b expression, whereas in the 7 7 B cell maturation other organs b7 expression was effectively repressed (Supplemental Fig. 2). This demonstrates that b7 indeed is the essential driving force The transcription factor SRF physically associates and synergizes for the homing to PP and MLn in these experiments. Interestingly, with KLF3, thereby working as a transactivation complex, as although b7-deficient B cells could effectively reach peripheral lymph shown for a muscle-specific target gene (12). Furthermore, mice in nodes, this was less pronounced for CD19:KLF3 transgenic B cells which SRF was deleted in B cells show a loss of MZ B cells (63), (Fig. 6A). a phenotype shared with KLF3-deficient mice (7, 9). Thus, these factors may act together to drive the formation of MZ B cells. To In b7-deficient mice, KLF3 still drives MZ B cell maturation investigate this possibility directly in vivo, the contribution of SRF by guest on September 29, 2021 Because KLF3 both repressed b7 and led to the accumulation of for MZ B cell maturation and b7 repression driven by KLF3 was MZ B cells in CD19:KLF3 transgenic mice, b7-deficient mice analyzed. As expected, in the absence of SRF, the MZ B cell were analyzed with respect to this B cell subset. As to this, b7 subset is significantly reduced (Fig. 8A [upper two dot plots], deficiency has no impact on the proportion of MZ B cells as de- 8B). MZ B cells, however, are rescued on the SRF-deficient

FIGURE 5. ChIP demonstrating KLF3 and KLF2 binding to the b7 integrin promoter. ChIP was performed on formaldehyde-fixed, sheared lysates of normal (white bars) and KLF3-transduced TK1 cells (gray bars, using the strep-tagII-myc-tag-KLF3 construct as in Fig. 4). For immunoprecipitation, either a goat polyclonal anti-KLF3 serum (A) or a rabbit polyclonal anti-KLF2 serum (B) was used. DNA recovered by immunoprecipitation was quantified with respect to the indicated target sequences by QRT-PCR. The percentage of input indicates the quantity of DNA corresponding to the given sample/PCR target sequence, as recovered following immunoprecipitation normalized relative to the input DNA quantity before immunoprecipitation. Results for two primer pairs amplifying nonoverlapping regions of the b7 integrin promoter are shown (Itgb7 sites 1 and 2). Control samples included those in which the antiserum was omitted (beads only control) or where the QRT-PCR was performed with a primer pair specific for an irrelevant genomic region. Data derive from duplicate determinations (shown are mean values and range). Results are representative of two experiments using independently generated lysates. 1742 KLF2/3-REGULATING LEUKOCYTE b7 INTEGRIN

results (Fig. 9). First, as expected for a KLF3-deficient (KLF3GCE/GCE) situation (9), b7 repression is impaired, such that in fraction F and among follicular B cells only 1.44% (0.25 + 1.19) and 2.73% (1.06 + 1.67) of the cells are b7neg as compared with 6.06% (5.9 + 0.16) and 8.11% (7.74 + 0.37) in normal mice, respectively, with KLF3GCE-heterozygous mice showing an intermediate result. Second, the presence of the CD19:KLF3 transgene leads to drastic repression of b7 in a significant fraction of the cells, such that neg .21% of the cells get b7 , whereas normally such cells are much less frequent. Third, although .90% of the cells in fraction F or in follicular B cells are GFP+ in KLF3GCE/GCE mice, the additional presence of the CD19:KLF3 transgene reduces this to ∼50% and below. The latter result is similar, although less obvious, for cells from KLF3GCE-heterozygous animals. Thus, whereas there is only a small population of b7neg cells among these cellular subsets normally, it appears that their presence depends on KLF3 repressing b7 expression, either by endogenous KLF3 or by the CD19: KLF3 transgene, with the latter being obviously more efficient.

Also, the transgene reduces GFP expression, both in KLF3GCE- Downloaded from heterozygous and -homozygous mice, agreeing well with the reduced endogenous KLF3 mRNA expression observed in the transgenic mice and being consistent with a negative feedback mechanism regulating KLF3 expression. Furthermore, GFP expression in KLF3GCE/GCE mice appears considerably higher than in the GCE KLF3 -heterozygous animals. Although in the heterozygous situ- http://www.jimmunol.org/ ation the remaining intact allele could be responsible for driving the feedback loop and thus limiting GFP expression, in KLF3GCE/GCE A FIGURE 6. KLF3 expression impacts lymphocyte homing. ( ) Shown mice this mechanism is abolished. However, in the latter case, are the ratios of Ly-5b and Ly-5a/b donor lymphocyte proportions among B + 2 a GFP expression also derives from two alleles. cells (CD19 93 ) obtained from the indicated organs of Ly-5 recipients b 15 h after adoptive transfer i.v. (B) Shows the corresponding data as in (A) Summarizing these data, KLF3 represses 7 and its own ex- for T cells (CD4+82 cells). (A and B) Data were obtained as shown in pression in vivo. However, even in the presence of the CD19: Supplemental Fig. 2. In three independent experiments, a total of four KLF3 transgene, the penetrance of this repression is signiﬁcant, recipient mice was analyzed for each of the three cellular mixtures but does not affect all cells. Thus, up to ∼40% of the cells show

b 2/2 by guest on September 29, 2021 encompassing B6, CD19:KLF3, and Itg 7 donor cells. Each of these repression of b7 and/or GFP, with the remaining unaffected pro- + + was an experimental group in all experiments. The dotted line indicates portion being b7 GFP . This may indicate that (a) further cofactor a 1:1 ratio, as expected from input numbers. Considering T or B cells, all (s), only expressed in a fraction of the cells, is/are essential for , , ratios 0.8 reached statistical significance relative to B6 cells (p 0.02). KLF3 to function as repressor. Because the above results indicate that KLF3 exerts a feedback background by the CD19:KLF3 transgene, demonstrating that the loop, ChIP was performed to determine whether there is direct interaction of SRF with KLF3 is dispensable for MZ B cell binding of KLF3 to the KLF3 promoter. Indeed, KLF3 binding maturation (Fig. 8A [lower left dot plot], 8B). Similarly, b7 re- to the KLF3 1b promoter, driving expression from the second pression by KLF3 is unaltered when SRF is absent (Fig. 8A noncoding exon [exon 2 in (57), exon 1b in (65)], is detected in [histogram plots], 8C). As in the normal situation, repression of b7 this study, either in normal or KLF3-overexpressing TK1 cells by KLF3 is more efficient among MZ B cells compared with their (Fig. 10A). Unfortunately, no suitable primer pair could be de- follicular counterpart. Thus, both KLF3-driven MZ B cell matu- rivedtodeterminebindingtotheexon1promoter,possiblydue . ration as well as b7 repression is SRF independent. to the region being GC-rich (partially 80% GC). In contrast to the b7 promoter, in which both KLF3 and KLF2 can bind, there Feedback regulation of KLF3 in vivo is no evidence by ChIP for KLF2 binding to the KLF3 promoter In CD19:KLF3 transgenic mice, endogenous KLF3 expression is (Fig. 10B). repressed as determined on mRNA and protein levels (7). To assay this effect more directly and to relate it to the repression of b7, Discussion the CD19:KLF3 transgene was combined with a newly available Target genes of the KLF transcription factor family have only KLF3 knock-in allele (Klf3-GCE). In this allele, a GFP-Cre-ERT2 partially been elucidated. In this study, we show that two factors of fusion gene construct is inserted into the first coding exon of this family, known to be relevant in lymphocytes, act as regulators KLF3, thereby inactivating KLF3 and allowing to monitor the of the lymphocyte-specific b7 integrin by directly binding to its KLF3 promoter activity by measuring GFP (45, 46) (http://www. promoter. The outcome is either increased expression or repres- gudmap.org/Docs/Mouse_Strains/Klf3.pdf). Analyzing the Hardy sion, as mediated by KLF2 or KLF3, respectively. This antago- fractions in the bone marrow [Supplemental Fig. 4, gating as to nistic outcome is reminiscent of the experimental result in vivo, in (64)], b7 starts to get appreciably expressed from fraction E on- which KLF3 is required for efficient B cell maturation toward the ward (Fig. 9A). In the presence of the Klf3-GCE allele, fraction E MZ B cell subset, whereas KLF2 limits this process (7, 9, 15, 16, is also the first to show evidence of GFP expression, particularly 19). Notably, it has been demonstrated that integrins mediate the when the allele is bred to homozygosity (compare fractions D and retention of MZ B cells within the MZ of the spleen by binding E of KLF3GCE/GCE; Fig. 9A). However, the analysis of bone ICAM-1 and VCAM-1 and, evidently for this purpose, MZ B cells marrow fraction F and peripheral B cells provided more clear express increased levels of LFA-1 (aL/b2) and a4/b1 integrin, The Journal of Immunology 1743

FIGURE 7. Follicular and MZ B cell subsets of b7 integrin-deﬁcient, CD19:KLF3 transgenic mice. (A) Splenic B cells of Itgb7-heterozygous and -homozygous mice were analyzed for CD1d, CD23, IgM, and CD21/35 to determine the presence and proportion of MZ B cells. Shown are CD19+CD932 cells. In total, n =9 Itgb7+/+ or +/2 and n = 8 Itgb72/2 animals were analyzed; one representative experiment is shown. (B) Splenic B cells of Itgb72/2 and Itgb72/2 CD19:KLF3 mice were analyzed for CD1d, CD23, and CD21/35. Shown are CD19+ CD932 cells. In total, n = 8 Itgb72/2 and n =3 Itgb72/2 CD19:KLF3 animals were analyzed; one representative experiment is shown. (C) Shown are the proportions of follicular and MZ B cells among CD19+CD932 cells as analyzed in (A) and (B). Data from Itgb7+/+ and Itgb7+/2 animals were pooled. The p values are indicated Downloaded from within the ﬁgure.

while having less b7, as compared with follicular B cells (66). heterodimers may preclude the experimental detection of relevant Accordingly, a possible hypothesis was that b7 repression by expression differences, once b7 is repressed. To specifically ad- http://www.jimmunol.org/ KLF3 allows for an augmented cell surface expression of a4/b1, dress the importance of b7 repression, b7-deficient mice were thus having higher affinity toward VCAM-1. This would increase ad- analyzed. No difference was detected as to the relative proportion hesion, and thus retention, within the MZ, thereby promoting MZ of follicular and MZ B cells, indicating that KLF3-mediated re- B cell survival and/or maturation. However, although b7 was pression of b7 is not the mechanism mediating the increase in MZ clearly downregulated, no change in the cell surface expression of B cells of CD19:KLF3 transgenic mice. Consequently, the CD19: a4 or b1 resulted, neither on MZ B cells of CD19:KLF3 transgenic KLF3 transgene remained effective in the promotion of MZ mice nor on KLF3-transduced P815 cells (data not shown), ar- B cells when bred onto the b7-deficient background. Thus, the guing against this simple scenario. Still, due to preferential pairing exchange of a4/b7 to a4/b1, as mediated by b7 repression, has no b a a b of 1 with 4 (37), small changes to the rather large pool of 4/ 1 relevance for the final share of MZ B cells to the B2B cell pool. by guest on September 29, 2021

FIGURE 8. Neither KLF3-mediated expansion of MZ B cells nor Itgb7 repression depends on the KLF3-interacting factor SRF. (A) Splenic B cells of SRFfl/fl mice, in which SRF was ablated in B cells by being CD19cre heterozygous, with or without the CD19:KLF3 transgene, were analyzed for CD21/35 and IgM to determine the proportion of MZ B cells. Shown are CD19+CD932 cells. Differentiation into follicular and MZ B cells is based on the gates, as shown on the left. Histograms depict the expression of Itgb7 among these subsets. Gray shades have been positioned into histograms at the same expression level to aid in the visualization of changes. A total of n = 4 SRFfl/fl CD19n/n, n = 5 SRFfl/fl CD19cre/n, and n = 5 SRFfl/fl CD19cre/n CD19:KLF3 transgenic mice was analyzed; one representative experiment is shown. (B and C) Shown are the proportions (B) and the geometric means and SDs for b7 expression (C) of follicular and MZ B cells, as analyzed in (A). The p values of relevant pairwise comparisons are indicated within the figure. 1744 KLF2/3-REGULATING LEUKOCYTE b7 INTEGRIN

FIGURE 9. A negative feedback loop controls KLF3 expression in vivo. (A) Bone marrow B cells of the indicated strains were analyzed for the expression of

GFP (reporting KLF3 expression) and b7 integrin. Shown are plots gated according to the Hardy-deﬁned scheme (64), corresponding to fractions D, E, and F (Supplemental Fig. 4). (B) Similar to (A), but for spleen. Gates were set on follicular B cells. (A and B) FACS data are from one representative experiment. The total number of analyzed mice is as follows: KLF3n/n (n =8),KLF3GCE/n (n = 11), KLF3GCE/n CD19:KLF3 (n =6),KLF3GCE/GCE (n = 4), and KLF3GCE/GCE CD19: KLF3 (n =3). Downloaded from http://www.jimmunol.org/

b7 integrin has been prominently known as an adhesion mole- still home to this region showed b7 expression. This affirms that cule mediating the homing and retention of lymphocytes to mu- b7 is the one relevant integrin controlled by KLF3 for homing to cosal sites (34). Indeed, b7-deficient mice were reported to have this side. Also, the occurrence of a small subset of cells in which smaller but normal numbers of PP (30). Accordingly, in CD19: KLF3 is unable to repress b7 explains why we could not detect KLF3 transgenic mice, fewer B cells would be expected to home a B cell paucity in these locations of CD19:KLF3 mice, unlike the to this site, due to the lack of b7 expression on many of them. We situation in animals lacking b7 throughout (30). Interestingly, by guest on September 29, 2021 compared T versus B cell proportions within PP of CD19:KLF3 CD19:KLF3 transgenic B cells fail to home efficiently to pe- transgenic and normal mice but found no difference (data not ripheral lymph nodes, but also do not accumulate in blood or shown). It is quite possible, though, that the population of B cells spleen, which is different from b7-deficient cells. This will require showing unaffected b7 expression in CD19:KLF3 transgenic mice further analyses. suffices to populate this tissue to normal numbers. To thus elu- As stated, not all B cells show b7 repression in CD19:KLF3 cidate the relevance of b7 repression by KLF3 in vivo, we transgenic mice. Cells that do not downregulate b7 were observed performed lymphocyte-homing assays, transferring mixtures of among bone marrow cells (within Hardy fractions E and F) and normal and CD19:KLF3 transgenic or b7-deficient cells. Indeed, follicular B cells, whereas most MZ B cells appear to be re- the outcome was that CD19:KLF3 transgenic B cells could not sponsive. Interestingly, the resistant cells are identical to those that efficiently reach PP and MLn, thus mimicking b7-deficient cells. also escaped the negative feedback loop when the CD19:KLF3 GCE + Interestingly, the few CD19:KLF3 transgenic B cells that could transgene was bred onto the KLF3 strain, appearing as b7

FIGURE 10. KLF3 binds its own promoter. (A and B) ChIP was performed as in Fig. 5 using normal and KLF3-transduced TK1 cells and either anti-KLF3 (A) or anti-KLF2 (B). Results for one primer pair amplifying a region of the KLF3 1b promoter as well as control regions are shown. Data derive from duplicate determinations (shown are mean values and range), and results are representative of two independent experiments using independently generated lysates. The Journal of Immunology 1745

GFP+ cells. This may indicate that either counteracting factors, 8. Turner, J., and M. Crossley. 1998. Cloning and characterization of mCtBP2, a co-repressor that associates with basic Kruppel-like€ factor and other mam- such as KLF2, may get even more upregulated at these stages, or malian transcriptional regulators. EMBO J. 17: 5129–5140. that there are further cofactors involved in the KLF3-mediated 9. Vu, T. T., D. Gatto, V. Turner, A. P. W. Funnell, K. S. Mak, L. J. Norton, W. Kaplan, M. J. Cowley, F. Agene`s, J. Kirberg, et al. 2011. Impaired B cell repression of b7 and in the feedback regulation of KLF3. The development in the absence of Kruppel-like€ factor 3. J. Immunol. 187: 5032– aforementioned B cell subsets would differ in the expression of 5042. these cofactors. Identified cofactors with which KLF3 interacts are 10. Takahashi, K., C. Matsumoto, and C. Ra. 2005. FHL3 negatively regulates hu- FHL3, CtBP2, and SRF (8, 11, 12). Some of these may indeed be man high-affinity IgE receptor beta-chain gene expression by acting as a transcriptional co-repressor of MZF-1. Biochem. J. 386: 191–200. relevant in lymphocytes, that is, SRF may act in conjunction with 11. Turner, J., H. Nicholas, D. Bishop, J. M. Matthews, and M. Crossley. 2003. The KLF3, as the absence of either of these leads to a loss/reduction of LIM protein FHL3 binds basic Kruppel-like€ factor/Kruppel-like€ factor 3 and its co-repressor C-terminal-binding protein 2. J. Biol. Chem. 278: 12786–12795. MZ B cells (7, 9, 63). Interaction with SRF may also be important 12. Himeda, C. L., J. A. Ranish, R. C. Pearson, M. Crossley, and S. D. Hauschka. for the repression of b7, particularly so because SRF has already 2010. KLF3 regulates muscle-specific gene expression and synergizes with se- been implicated in hematopoietic stem cell adhesion by regulating rum response factor on KLF binding sites. Mol. Cell. Biol. 30: 3430–3443. 13. Bai, A., H. Hu, M. Yeung, and J. Chen. 2007. Kruppel-like factor 2 controls integrin expression (67). With respect to the latter, however, nei- T cell trafficking by activating L-selectin (CD62L) and sphingosine-1-phosphate ther the interaction of KLF3 with CtBP2 nor with SRF appears to receptor 1 transcription. J. Immunol. 178: 7632–7639. be relevant in this context, because b repression ensued even 14. Carlson, C. M., B. T. Endrizzi, J. Wu, X. Ding, M. A. Weinreich, E. R. Walsh, 7 M. A. Wani, J. B. Lingrel, K. A. Hogquist, and S. C. Jameson. 2006. Kruppel- when a CtBP2 interaction motif–defective form of KLF3 was like factor 2 regulates thymocyte and T-cell migration. Nature 442: 299–302. expressedinP815cellsorwhenKLF3wasexpressedinvivoin 15. Hart, G. T., X. Wang, K. A. Hogquist, and S. C. Jameson. 2011. Kruppel-like€ factor 2 (KLF2) regulates B-cell reactivity, subset differentiation, and trafficking B cells lacking SRF. Also, in the absence of SRF, the MZ B cell molecule expression. Proc. Natl. Acad. Sci. USA 108: 716–721. pool was still expanded by constitutive KLF3 expression. This 16. Hoek, K. L., L. E. Gordy, P. L. Collins, V. V. Parekh, T. M. Aune, S. Joyce, Downloaded from indicates that KLF3 either acts downstream of SRF or in an- J. W. Thomas, L. Van Kaer, and E. Sebzda. 2010. Follicular B cell trafficking within the spleen actively restricts humoral immune responses. Immunity 33: other pathway, balancing SRF deficiency. 254–265. An alternative explanation for the lack of uniform b7 repression 17. Schober, S. L., C. T. Kuo, K. S. Schluns, L. Lefrancois, J. M. Leiden, and might be that KLF3 repression only affects cells having a specific S. C. Jameson. 1999. Expression of the transcription factor lung Kruppel-like€ factor is regulated by cytokines and correlates with survival of memory T cells migratory history or destiny, such that KLF3-mediated repression in vitro and in vivo. J. Immunol. 163: 3662–3667. would be transient. Such a mechanism could be relevant for the 18. Weinreich, M. A., K. Takada, C. Skon, S. L. Reiner, S. C. Jameson, and http://www.jimmunol.org/ K. A. Hogquist. 2009. KLF2 transcription-factor deficiency in T cells results in recirculation of lymphocytes or the exclusion from this process. unrestrained cytokine production and upregulation of bystander chemokine For instance, MZ B cells have been noted to be particularly sta- receptors. Immunity 31: 122–130. tionary (7, 68). 19. Winkelmann, R., L. Sandrock, M. Porstner, E. Roth, M. Mathews, E. Hobeika, M. Reth, M. L. Kahn, W. Schuh, and H. M. Ja¨ck. 2011. B cell homeostasis and Finally, it is shown in this work that KLF3 performs an efficient plasma cell homing controlled by Kruppel-like€ factor 2. Proc. Natl. Acad. Sci. negative regulatory feedback loop by binding to and repressing its USA 108: 710–715. own promoter. This feature appears to be a shared trait among KLFs 20. Wu, J., and J. B. Lingrel. 2005. Kruppel-like€ factor 2, a novel immediate-early transcriptional factor, regulates IL-2 expression in T lymphocyte activation. J. because a similar regulatory loop was described for KLF4 (GKLF) Immunol. 175: 3060–3066. (69). KLF4 also has been implicated in the regulation of integrins, 21. Dang, X., N. A. Raffler, and K. Ley. 2009. Transcriptional regulation of mouse

L-selectin. Biochim. Biophys. Acta 1789: 146–152. by guest on September 29, 2021 because among myeloid cells KLF4 represses CD11d expression 22. Sebzda, E., Z. Zou, J. S. Lee, T. Wang, and M. L. Kahn. 2008. Transcription (70), an integrin (aD) specifically expressed by red pulp macro- factor KLF2 regulates the migration of naive T cells by restricting chemokine phages (71). Thus, many evidences point to a general involvement receptor expression patterns. Nat. Immunol. 9: 292–300. 23. Takada, K., X. Wang, G. T. Hart, O. A. Odumade, M. A. Weinreich, of KLFs in the positioning of lymphocytes, particularly by regu- K. A. Hogquist, and S. C. Jameson. 2011. Kruppel-like factor 2 is required for lating integrin expression. trafficking but not quiescence in postactivated T cells. J. Immunol. 186: 775–783. 24. Weinreich, M. A., and K. A. Hogquist. 2008. Thymic emigration: when and how T cells leave home. J. Immunol. 181: 2265–2270. Acknowledgments 25. Butcher, E. C., and L. J. Picker. 1996. Lymphocyte homing and homeostasis. We thank Y. Joos, D. Sprugel,€ and I. Grossmann for expert technical assis- Science 272: 60–66. 26. Pauls, K., M. Schon, R. C. Kubitza, B. Homey, A. Wiesenborn, P. Lehmann, tance; the animal caretakers for help; Merlin Crossley for the pMT2/BKLF ¨ T. Ruzicka, C. M. Parker, and M. P. Scho¨n. 2001. Role of integrin alphaE plasmid; Elisabeth Kremmer for rat anti-KLF3 mAb 511.7D12 supernatant; (CD103)beta7 for tissue-specific epidermal localization of CD8+ T lymphocytes. and Sebastian Newrzela for TK-1 cells. J. Invest. Dermatol. 117: 569–575. 27. Peters, T., W. Bloch, O. Pabst, C. Wickenhauser, C. Uthoff-Hachenberg, S. V. Schmidt, G. Varga, S. Grabbe, D. Kess, T. Oreshkova, et al. 2012. Adaptive Disclosures immune response to model antigens is impaired in murine leukocyte-adhesion The authors have no financial conflicts of interest. deficiency-1 revealing elevated activation thresholds in vivo. Clin. Dev. Immunol. DOI: 10.1155/2012/450738. 28. Scharffetter-Kochanek, K., H. Lu, K. Norman, N. van Nood, F. Munoz, S. Grabbe, M. McArthur, I. Lorenzo, S. Kaplan, K. Ley, et al. 1998. Spontaneous References skin ulceration and defective T cell function in CD18 null mice. J. Exp. Med. 1. van Vliet, J., L. A. Crofts, K. G. Quinlan, R. Czolij, A. C. Perkins, and 188: 119–131. M. Crossley. 2006. Human KLF17 is a new member of the Sp/KLF family of 29. Steeber, D. A., M. L. Tang, X. Q. Zhang, W. Muller,€ N. Wagner, and T. F. Tedder. transcription factors. Genomics 87: 474–482. 1998. Efficient lymphocyte migration across high endothelial venules of mouse 2. Feinberg, M. W., Z. Lin, S. Fisch, and M. K. Jain. 2004. An emerging role for Peyer’s patches requires overlapping expression of L-selectin and beta7 integrin. Kruppel-like€ factors in vascular biology. Trends Cardiovasc. Med. 14: 241–246. J. Immunol. 161: 6638–6647. 3. Kaczynski, J., T. Cook, and R. Urrutia. 2003. Sp1- and Kruppel-like€ transcription 30. Wagner, N., J. Lo¨hler, E. J. Kunkel, K. Ley, E. Leung, G. Krissansen, factors. Genome Biol. 4: 206. K. Rajewsky, and W. Muller.€ 1996. Critical role for beta7 integrins in formation 4. Lomberk, G., and R. Urrutia. 2005. The family feud: turning off Sp1 by Sp1-like of the gut-associated lymphoid tissue. Nature 382: 366–370. KLF proteins. Biochem. J. 392: 1–11. 31. Wagner, N., J. Lo¨hler, T. F. Tedder, K. Rajewsky, W. Muller,€ and D. A. Steeber. 5. Turner, J., and M. Crossley. 1999. Mammalian Kruppel-like€ transcription fac- 1998. L-selectin and beta7 integrin synergistically mediate lymphocyte migra- tors: more than just a pretty finger. Trends Biochem. Sci. 24: 236–240. tion to mesenteric lymph nodes. Eur. J. Immunol. 28: 3832–3839. 6. Crossley, M., E. Whitelaw, A. Perkins, G. Williams, Y. Fujiwara, and 32. La¨mmermann, T., B. L. Bader, S. J. Monkley, T. Worbs, R. Wedlich-So¨ldner, S. H. Orkin. 1996. Isolation and characterization of the cDNA encoding BKLF/ K. Hirsch, M. Keller, R. Fo¨rster, D. R. Critchley, R. Fa¨ssler, and M. Sixt. 2008. TEF-2, a major CACCC-box-binding protein in erythroid cells and selected Rapid leukocyte migration by integrin-independent flowing and squeezing. other cells. Mol. Cell. Biol. 16: 1695–1705. Nature 453: 51–55. 7. Turchinovich, G., T. T. Vu, F. Frommer, J. Kranich, S. Schmid, M. Alles, 33. Schippers, A., C. Leuker, O. Pabst, A. Kochut, B. Prochnow, A. D. Gruber, J.-B. Loubert, J.-P. Goulet, U. Zimber-Strobl, P. Schneider, et al. 2011. E. Leung, G. W. Krissansen, N. Wagner, and W. Muller.€ 2009. Mucosal addressin Programming of marginal zone B-cell fate by basic Kruppel-like factor (BKLF/ cell-adhesion molecule-1 controls plasma-cell migration and function in the KLF3). Blood 117: 3780–3792. small intestine of mice. Gastroenterology 137: 924–933. 1746 KLF2/3-REGULATING LEUKOCYTE b7 INTEGRIN

34. Berlin, C., E. L. Berg, M. J. Briskin, D. P. Andrew, P. J. Kilshaw, B. Holzmann, cell development by the dominant negative helix loop helix factor Id3. J. Exp. I. L. Weissman, A. Hamann, and E. C. Butcher. 1993. Alpha 4 beta 7 integrin Med. 186: 1597–1602. mediates lymphocyte binding to the mucosal vascular addressin MAdCAM-1. 52. Schlaeger, E. J. 1996. The protein hydrolysate, Primatone RL, is a cost-effective Cell 74: 185–195. multiple growth promoter of mammalian cell culture in serum-containing and 35. Day, E. S., L. Osborn, and A. Whitty. 2002. Effect of divalent cations on the serum-free media and displays anti-apoptosis properties. J. Immunol. Methods affinity and selectivity of alpha4 integrins towards the integrin ligands vascular 194: 191–199. cell adhesion molecule-1 and mucosal addressin cell adhesion molecule-1: Ca2+ 53. Morita, S., T. Kojima, and T. Kitamura. 2000. Plat-E: an efficient and stable activation of integrin alpha4beta1 confers a distinct ligand specificity. Cell system for transient packaging of retroviruses. Gene Ther. 7: 1063–1066. Commun. Adhes. 9: 205–219. 54. Luthman, H., and G. Magnusson. 1983. High efficiency polyoma DNA trans- 36. Luster, A. D., R. Alon, and U. H. von Andrian. 2005. Immune cell migration in fection of chloroquine treated cells. Nucleic Acids Res. 11: 1295–1308. inflammation: present and future therapeutic targets. Nat. Immunol. 6: 1182– 55. Graham, F. L., and A. J. van der Eb. 1973. A new technique for the assay of 1190. infectivity of human adenovirus 5 DNA. Virology 52: 456–467. 37. DeNucci, C. C., A. J. Pagań, J. S. Mitchell, and Y. Shimizu. 2010. Control of 56. Schuh, W., S. Meister, K. Herrmann, H. Bradl, and H. M. Ja¨ck. 2008. Tran- alpha4beta7 integrin expression and CD4 T cell homing by the beta1 integrin scriptome analysis in primary B lymphoid precursors following induction of the subunit. J. Immunol. 184: 2458–2467. pre-B cell receptor. Mol. Immunol. 45: 362–375. 38. Chen, H. M., H. L. Pahl, R. J. Scheibe, D. E. Zhang, and D. G. Tenen. 1993. The 57. Kirberg,J.,C.Gschwendner,J.P.Dangy,F.Ruckerl,€ F. Frommer, and J. Bachl. 2005. Sp1 transcription factor binds the CD11b promoter specifically in myeloid cells Proviral integration of an Abelson-murine leukemia virus deregulates BKLF- in vivo and is essential for myeloid-specific promoter activity. J. Biol. Chem. expression in the hypermutating pre-B cell line 18-81. Mol. Immunol. 42: 1235–1242. 268: 8230–8239. 58. Ralph, P., and I. Nakoinz. 1974. Lipopolysaccharides inhibit lymphosarcoma 39. Gaudreault, M., F. Vigneault, S. Leclerc, and S. L. Gue´rin. 2007. Laminin cells of bone marrow origin. Nature 249: 49–51. reduces expression of the human alpha6 integrin subunit gene by altering the 59. Butcher, E. C., R. G. Scollay, and I. L. Weissman. 1980. Organ specificity level of the transcription factors Sp1 and Sp3. Invest. Ophthalmol. Vis. Sci. 48: of lymphocyte migration: mediation by highly selective lymphocyte interaction 3490–3505. with organ-specific determinants on high endothelial venules. Eur. J. Immunol. 40. Gingras, M.-E., B. Masson-Gadais, K. Zaniolo, S. Leclerc, R. Drouin, 10: 556–561. L. Germain, and S. L. Gue´rin. 2009. Differential binding of the transcription 60. Holzmann, B., and I. L. Weissman. 1989. Peyer’s patch-specific lymphocyte factors Sp1, AP-1, and NFI to the promoter of the human alpha5 integrin gene homing receptors consist of a VLA-4-like alpha chain associated with either Downloaded from dictates its transcriptional activity. Invest. Ophthalmol. Vis. Sci. 50: 57–67. of two integrin beta chains, one of which is novel. EMBO J. 8: 1735–1741. 41. Markovics, J. A., J. Araya, S. Cambier, D. Jablons, A. Hill, P. J. Wolters, and 61. Holzmann, B., B. W. McIntyre, and I. L. Weissman. 1989. Identification of a mu- S. L. Nishimura. 2010. Transcription of the transforming growth factor beta rine Peyer’s patch—specific lymphocyte homing receptor as an integrin molecule activating integrin beta8 subunit is regulated by SP3, AP-1, and the p38 pathway. with an alpha chain homologous to human VLA-4 alpha. Cell 56: 37–46. J. Biol. Chem. 285: 24695–24706. 62. Kilshaw, P. J., and S. J. Murant. 1991. Expression and regulation of beta 7(beta p) 42. Rosmarin, A. G., M. Luo, D. G. Caprio, J. Shang, and C. P. Simkevich. 1998. integrins on mouse lymphocytes: relevance to the mucosal immune system. Eur. Sp1 cooperates with the ets transcription factor, GABP, to activate the CD18 J. Immunol. 21: 2591–2597. € (beta2 leukocyte integrin) promoter. J. Biol. Chem. 273: 13097–13103. 63. Fleige, A., S. Alberti, L. Gro¨be, U. Frischmann, R. Geffers, W. Muller, http://www.jimmunol.org/ 43. Leung, E., P. E. Mead, Q. Yuan, W. M. Jiang, J. D. Watson, and A. Nordheim, and A. Schippers. 2007. Serum response factor contributes se- G. W. Krissansen. 1993. The mouse beta 7 integrin gene promoter: transcrip- lectively to lymphocyte development. J. Biol. Chem. 282: 24320–24328. tional regulation of the leukocyte integrins LPAM-1 and M290. Int. Immunol. 5: 64. Hardy, R. R., C. E. Carmack, S. A. Shinton, J. D. Kemp, and K. Hayakawa. 551–558. 1991. Resolution and characterization of pro-B and pre-pro-B cell stages in 44. Lim, S. P., E. Leung, and G. W. Krissansen. 1998. The beta7 integrin gene (Itgb- normal mouse bone marrow. J. Exp. Med. 173: 1213–1225. 7) promoter is responsive to TGF-beta1: defining control regions. Immunoge- 65. Funnell, A. P., C. A. Maloney, L. J. Thompson, J. Keys, M. Tallack, netics 48: 184–195. A. C. Perkins, and M. Crossley. 2007. Erythroid Kruppel-like€ factor directly 45. Harding, S. D., C. Armit, J. Armstrong, J. Brennan, Y. Cheng, B. Haggarty, activates the basic Kruppel-like€ factor gene in erythroid cells. Mol. Cell. Biol. D. Houghton, S. Lloyd-MacGilp, X. Pi, Y. Roochun, et al. 2011. The GUDMAP 27: 2777–2790. database—an online resource for genitourinary research. Development 138: 66. Lu, T. T., and J. G. Cyster. 2002. Integrin-mediated long-term B cell retention in 2845–2853. the splenic marginal zone. Science 297: 409–412.

46. McMahon, A. P., B. J. Aronow, D. R. Davidson, J. A. Davies, K. W. Gaido, 67. Ragu, C., G. Elain, E. Mylonas, C. Ottolenghi, N. Cagnard, D. Daegelen, by guest on September 29, 2021 S. Grimmond, J. L. Lessard, M. H. Little, S. S. Potter, E. L. Wilder, et al. 2008. E. Passegue´, W. Vainchenker, O. A. Bernard, and V. Penard-Lacronique. 2010. GUDMAP: the genitourinary developmental molecular anatomy project. J. Am. The transcription factor Srf regulates hematopoietic stem cell adhesion. Blood Soc. Nephrol. 19: 667–671. 116: 4464–4473. 47. Miano, J. M., N. Ramanan, M. A. Georger, K. L. de Mesy Bentley, 68. Gray, D., I. C. MacLennan, H. Bazin, and M. Khan. 1982. Migrant mu+ delta+ R. L. Emerson, R. O. Balza, Jr., Q. Xiao, H. Weiler, D. D. Ginty, and R. P. Misra. and static mu+ delta2 B lymphocyte subsets. Eur. J. Immunol. 12: 564–569. 2004. Restricted inactivation of serum response factor to the cardiovascular 69. Mahatan, C. S., K. H. Kaestner, D. E. Geiman, and V. W. Yang. 1999. Char- system. Proc. Natl. Acad. Sci. USA 101: 17132–17137. acterization of the structure and regulation of the murine gene encoding gut- 48. Rickert, R. C., K. Rajewsky, and J. Roes. 1995. Impairment of T-cell-dependent enriched Kruppel-l€ ike factor (Kruppel-like€ factor 4). Nucleic Acids Res. 27: B-cell responses and B-1 cell development in CD19-deficient mice. Nature 376: 4562–4569. 352–355. 70. Noti, J. D., A. K. Johnson, and J. D. Dillon. 2005. The leukocyte integrin gene 49. Rickert, R. C., J. Roes, and K. Rajewsky. 1997. B lymphocyte-specific, Cre- CD11d is repressed by gut-enriched Kruppel-like factor 4 in myeloid cells. J. mediated mutagenesis in mice. Nucleic Acids Res. 25: 1317–1318. Biol. Chem. 280: 3449–3457. 50. Kitamura, T., Y. Koshino, F. Shibata, T. Oki, H. Nakajima, T. Nosaka, and 71. Miyazaki, Y., M. Bunting, D. M. Stafforini, E. S. Harris, T. M. McIntyre, H. Kumagai. 2003. Retrovirus-mediated gene transfer and expression cloning: S. M. Prescott, V. S. Frutuoso, F. C. Amendoeira, D. de Oliveira Nascimento, powerful tools in functional genomics. Exp. Hematol. 31: 1007–1014. A. Vieira-de-Abreu, et al. 2008. Integrin alphaDbeta2 is dynamically expressed 51. Heemskerk, M. H., B. Blom, G. Nolan, A. P. Stegmann, A. Q. Bakker, K. Weijer, by inflamed macrophages and alters the natural history of lethal systemic P. C. Res, and H. Spits. 1997. Inhibition of T cell and promotion of natural killer infections. J. Immunol. 180: 590–600.