Published online before print April 28, 2009
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* Instituto de Investigaciones Biotecnológicas-Instituto Tecnológico de Chascomús, CONICET-Universidad Nacional de San Martín, Buenos Aires, Argentina; and
The Ludwig Institute for Cancer Research, Uppsala University, Sweden
1. Correspondence: Instituto de Investigaciones Biotecnológicas, Av General Paz 5445, Predio INTI, Edificio 24, B1650WBA San Martin, Buenos Aires, Argentina. E-mail: oscar{at}unsam.edu.ar
ABSTRACT
Galectin (Gal) constitute a family of carbohydrate-recognizing molecules ubiquitously expressed in mammals. In the immune system, they regulate many processes such as inflammation, adhesion, and apoptosis. Here, we report the expression in the spleen of the two same Gal-8 splice variants described previously in the thymus. Gal-8 was found to induce two separate biological activities on T lymphocytes: a robust naive CD4+ T cell proliferation in the absence of antigen and notably, a costimulatory signal that synergized the cognate OVA peptide in DO11.10 mice transgenic for TCROVA. The antigen-independent proliferation induced by Gal-8 displayed increased expression of pro- and anti-inflammatory cytokines, thus suggesting the polyclonal expansion of Th1 and Th2 clones. The costimulatory effect on antigen-specific T cell activation was evidenced when the Gal and the peptide were assayed at doses suboptimal to induce T cell proliferation. By mass spectra analysis, several integrins and leukocyte surface markers, including CD45 isoforms, as well as other molecules specific to macrophages, neutrophils, and platelets, were identified as putative Gal-8 counter-receptors. Gal-8 triggered pZAP70 and pERK1/2. Moreover, pretreatment with specific inhibitors of CD45 phosphatase or ERK1/2 prevented its antigen-dependent and -independent T cell-proliferative activities. This seems to be associated with the agonistic binding to CD45, which lowers the activation threshold of the TCR signaling pathway. Taken together, our findings support a distinctive role for locally produced Gal-8 as an enhancer of otherwise borderline immune responses and also suggest that Gal-8 might fuel the reactivity at inflammatory foci.
Key Words: cell surface molecules • cell activation • CD45 • integrins
Introduction
Protein glycosylation is becoming a rapidly growing study field in cell and tissue physiology as a result of the huge complexity involved that allows a fine regulation of several processes [1 ]. In fact, the interaction between different cell types and their environment relies in part on their differential glycosylation profiling. Together with many surface molecules able to recognize glycosylated binders, some soluble mediators such as Gal are involved in homeostasis and in pathological events [2 , 3 ]. Gal constitute a family of β-galactoside binding mammalian lectins that contains conserved carbohydrate recognition domains. They are classified in three groups based on their structure: prototype such as Gal-1; chimera type with Gal-3 as its only representative; and tandem repeat type, where Gal-8 is included. As they are involved in a growing number of different biological processes, their structure and biochemistry are currently receiving strong attention.
Regarding their activity on the immune system, several members of this protein family including Gal-1, Gal-3, Gal-8, and Gal-9 [2 , 4 5 6 7 ] are expressed by different cell populations, where acting mostly in an autocrine/paracrine manner, they are involved in several phenomena such as regulation, adhesion, differentiation, and apoptosis [8 9 10 ]. Some of them are also associated with tumor evasion and progression [3 ]. The activity of some Gal, particularly Gal-1 and Gal-3, in the immune system has been analyzed in detail, but in contrast, scarce information about the effects of Gal-8 is still available. Gal-8 belongs to the tandem repeat-type group, which contains two carbohydrate recognition domains joined by a linker peptide, whose variable length defines different isoforms in mice, rats, and humans [7 , 11 12 13 ]. Gal-8 is widely expressed in several tissues under normal and altered conditions [2 ]. Through the interaction with integrins [13 , 14 ], Gal-8 induces an adhesive phenotype in several cells such as Jurkat T cells [15 , 16 ] and neutrophils [17 ]. Moreover, Gal-8-induced spreading of Jurkat cells depends on the interaction of this Gal with specific β1 integrins, thus suggesting a modulating property in the cell-driven immune response [15 ]. We have shown recently that two isoforms of Gal-8 differing in the length of their linker peptide are present in the thymus, where both are able to induce apoptosis of the CD4highCD8high thymocytes through a mechanism involving the caspase activation pathway [7 ]. The aim of the present work was to analyze the expression of Gal-8 in the spleen and explore its functions on peripheral lymphocytes.
MATERIALS AND METHODS
Mice, cell lines, and cell purification
C57BL/6J and C.Cg-Tg (DO11.10) 10Dlo/J (DO11.10) breeding pairs were obtained from The Jackson Laboratories (Bar Harbor, ME, USA) and bred in our animal facilities. The Committee of Ethics of the Instituto de Investigaciones Biotecnológicas (Buenos Aires, Argentina) approved all animal experiments. Dr. Anne I. Sperling (University of Chicago, Chicago, IL, USA) generously provided B6.CD43null mice, backcrossed eight times to C57BL/6J mice. Athymic NIHnu/nu or euthymic NIH+/+ mice were obtained from the animal facility of the National University of La Plata (Buenos Aires, Argentina). The human Jurkat T cell line was obtained from American Type Culture Collection (Manassas, VA, USA). For mouse splenocyte purification, spleens from 4- to 8-week-old animals were removed and disrupted against a stainless steel mesh in RPMI-1640 medium (Invitrogen, Carlsbad, CA, USA). The cell suspension was washed and incubated with RBC lysis buffer (Sigma Chemical Co., St. Louis, MO, USA) and washed again with medium. For T cell purification, C57BL/6J mouse splenocytes were treated with anti-CD19-coated beads (Dako, Carpinteria, CA, USA) following the manufacturer’s protocol for cell depletion. Purity was confirmed by flow cytometry with PE-labeled anti-CD19 (Becton Dickinson, San Jose, CA, USA). CD4 and CD8 T cells were purified from splenocytes from C57BL/6J mice by using the miniMACS columns and anti-CD4- or anti-CD8-coupled paramagnetic particles from Miltenyi Biotec (Germany). Cells assayed by FACS for purity indicated no contaminants in the eluted population. Cell lines and mouse primary cells were cultured at 37°C in 5% CO2 in RPMI-1640 in the presence of 10% FBS (Invitrogen), 2 mM glutamine, and gentamycin (complete medium).
Antibodies and reagents
Anti-mouse Gal-8 rabbit antibodies affinity-purified through a Gal-8-Sepharose column were obtained as described earlier [7
]. Dr. Yehiel Zick (The Weizmann Institute of Science, Israel) kindly provided a mouse anti-rat Gal-8 mAb [18
]. Fluorescent mAb against mouse CD19 was from Becton Dickinson. The anti-p44/42 MAPK (ERK1/2), pp44/42 MAPK (Thr202/Tyr204), ZAP70, and pZAP70 antibodies and the U0126 ERK1/2 phosphorylation inhibitor were from Cell Signaling Technology (Beverly, MA, USA). The CD45 PTPase inhibitor N-(9, 10-dioxo-9,10-dihydro-phenanthren-2-yl)-2,2-dimethyl-propionamide [19
] was from Calbiochem (La Jolla, CA, USA). TDG was from Sigma Chemical Co.
Expression of mouse rGal-8
Both recombinant isoforms of Gal-8 cloned from mouse thymic cells (Gal-8L and Gal-8S) were expressed in Escherichia coli and purified by two steps of affinity chromatography as described previously [7
]. Lectin activity of these molecules was tested by hemagglutination assays [7
].
RT-PCR
Splenocytes were solubilized in Trizol reagent (Invitrogen), and total RNA was extracted following the manufacturer’s instructions. mRNA was purified using the PolyATract mRNA isolation kit from Promega (Madison, WI, USA) and used as template for the retrotranscriptase Superscript II (Invitrogen) to obtain cDNA. Controls without the retrotranscriptase were included. Gal-8 transcripts were detected with a set of primers (forward 5'-AAAGCTAGCATGTTGTCCTTAAATAACCTA-3', reverse 5'-AAAAGATCTCCGGATAAGCCATTTTGTA-3'), which amplifies the entire coding sequence of Gal-8 using mouse splenocyte cDNA as template.
Western blot
Splenocyte lysates from C57BL/6J male mice were seeded onto a lactosyl-Sepharose column (Sigma Chemical Co.) and eluted with lactose as described [7
]. Eluates were solved in a 10% SDS-PAGE and transferred to polyvinylidene difluoride membranes. Blots were probed with affinity-purified anti-mouse Gal-8 antibodies 1/100, followed by HRP-labeled goat IgG anti-rabbit IgG 1/5000, and developed by chemiluminescence (both from Pierce, Rockford, IL, USA). In a similar set of assays, blots were probed with a mouse mAb anti-rat Gal-8 followed by HRP-labeled anti-mouse IgG secondary antibodies (Dako).
Gal-8 binding assays
Mouse C57BL/6J splenocytes were treated with Gal-8, 0.1 µM, washed with cold PBS or PBS plus 100 mM lactose, blocked with anti-FcRs (Becton Dickinson), and incubated with affinity-purified rabbit anti-Gal-8 antibodies [7
] with sodium azide in an ice bath, followed by incubation with 1/1000 FITC-conjugated anti-rabbit IgG antibodies (Molecular Probes, Invitrogen, Eugene, Oregon, USA). Cells were washed, fixed, and analyzed by flow cytometry. Two controls were included: one where cells were first incubated with a nonrelated rabbit IgG and another one where cells were processed without incubation with Gal-8. A FACSCalibur cytometer (Becton Dickinson) and WinMdi software were used throughout this work.
Proliferation and costimulation assays
Splenocytes (5x105) from C57BL/6J male mice were cultured for 48 h in 96-multiwell plates in a final volume of 0.2 ml RPMI-1640 complete medium. An aliquot of 50 µl containing 1 µCi [3H]-methyl thymidine (New England Nuclear, Boston, MA, USA) was added to each well for the last 16 h before cell harvesting. TDG or CD45 PTPase inhibitor was always added 20 min before Gal-8. U0126 inhibitor (10 µM) was added 1 h before the assay. No significant differences with basal cpm were found in TDG, PTPase, or U0126 inhibitors, only added cultures. An aliquot of 5 µg/ml LPS or Con A (Sigma Chemical Co.) was added as proliferation controls when indicated. For costimulatory assays, splenocytes (1.5x105 cells) were taken from 4- to 8-week-old female DO11.10 mice and cultured in a final volume of 0.1 ml in the presence of 0.4–1 µg/ml cognate OVA323–339 peptide (Sigma-Genosys, The Woodlands, TX, USA) for 48 h, adding [3H]-methyl thymidine 16 h before harvesting. For short assays, 24 h cultures were carried out with the addition of labeled thymidine 6 h before collection. The peptide dose was determined in preliminary experiments (not shown) so as to induce low cell proliferation under these conditions. All assays were performed in quadruplicate.
Affinity chromatography and mass spectrometry
Splenocytes from C75BL/6J mice were lysed with 1% Triton X-100 in the presence of a cocktail of protease inhibitors and passed through a Gal-8 affinity column made by coupling purified rGal-8L to N-hydroxysciccinimide-activated HiTrap columns (GE Healthcare, Uppsala, Sweden). Gal-8 binders were eluted with 100 mM lactose in PBS and concentrated by Amicon Centriprep YM-3 (Millipore, Bedford, MA, USA). After alkylation with iodoacetamide, samples were resuspended in cracking buffer and solved by 10% SDS-PAGE. After Coomassie blue staining, discrete bands were cut out, in-gel digested overnight with trypsin, and subjected to peptide mass fingerprinting on an Ultraflex TOF/TOF (Bruker Daltonics, Bremen, Germany) MALDI mass spectrometer.
RT-MPCR
For RT-MPCR assays, the MPCR Kit for Mouse Th1/Th2 Cytokines Set-2 (Maxims Biotech, Rockville, MD, USA) was used. This kit contains specific primers for IFN-
, IL-2, IL-4, IL-5, IL-10, IL-12, IL-13, and GAPDH as an internal control. cDNA from splenocytes cultured for 4–24 h in the presence of Gal-8 or Gal-8 plus TDG was used as template for these assays. Gal-8-treated splenocytes were washed with 200 mM lactose before harvesting to detach them from the plate.
pp44/42 MAPK (PERK1/2)
Jurkat cells (3x106) were cultured in complete medium for 1 h in the presence of Gal-8, alone or in combination with TDG or CD45 PTPase inhibitor. Cells were washed with cold PBS plus lactose 100 mM before harvesting to detach them from the plate bottom and then washed with cold PBS. Western blots were carried out following the manufacturer’s protocol (Cell Signaling Technology), with the addition of 2 mM sodium orthovanadate to the lysis buffer as the only modification.
pZAP70
Splenocytes were obtained from DO11.10 female mice, and 5 x 106 nucleated cells were cultured for 1 h in a final volume of 2 ml RPMI complete medium with the addition of 0.1 µM Gal-8, 1 µg/ml OVA323–339 peptide, and 2 µM CD45 PTPase inhibitor as indicated and processed as for PERK1/2.
Statistical analysis
Student’s t-test was used.
RESULTS
Expression of Gal-8 in the mouse spleen
Gal-8 transcripts were detected by RT-PCR using mouse splenocyte cDNA as template (Fig. 1A
), and the main amplicon (
1 kbp band) was cloned. A control without the retrotranscriptase rendered no signal. Noticeably, the two same isoforms (Gal-8S and Gal-8L), found previously in thymocytes [7
], were identified in mouse splenocytes. Nine independent clones were sequenced that resulted in eight corresponding to Gal-8S and one to Gal8-L. To confirm the actual presence of the protein, mouse spleen homogenates were loaded onto lactosyl-Sepharose affinity columns, and the fraction eluted with lactose was probed by Western blot using affinity-purified anti-Gal-8 rabbit polyclonal antibodies and a mouse anti-rat Gal-8 mAb [18
] (Fig. 1B)
.
 View larger version (47K): [in a new window] |
Figure 1. Expression of Gal-8 in mouse spleen. (A) RT-PCR. cDNA from splenocytes and thymocytes were used as template in a RT-PCR assay with specific primers for mouse Gal-8 open-reading frame, which results in 1 kb-length transcripts. (B) Western blots. Whole homogenates from spleens or thymuses (as positive control) were affinity-purified by lactosyl-Sepharose, and the eluted fractions were precipitated with trichloroacetic acid and subjected to 10% SDS-PAGE. Blots were reacted with affinity-purified rabbit anti-mouse Gal-8 antibodies or mouse anti-rat Gal-8 mAb. A lower molecular weight band in the monoclonal-developed Western blot corresponds to mouse Igs.
|
 View larger version (16K): [in a new window] |
Figure 2. Gal-8 induces naïve T cell proliferation. (A) Gal-8 binding to splenocyte surface. Splenocytes were treated with 0.1 µM Gal-8L and binding tested with rabbit anti-Gal-8 antibodies. Control, Anti-Gal-8 antibodies in the absence of Gal-8 addition; Lac, cells washed with 100 mM lactose after Gal-8 treatment; Normal IgG, antibodies purified from naive rabbits. (B) Splenocytes were cultured with the indicated Gal-8L amounts and proliferation quantified. TDG was added at 30 mM 20 min before Gal-8 addition. No significant differences with basal cpm were found in cultures where only TDG was added. *, P < 0.01; **, P < 0.001. Results are representative of more than five independent assays.
|
CD4 T cells are the main targets for Gal-8-induced proliferation
To narrow down the population of Gal-8-responsive cells in the proliferation studies, assays similar to those performed before were carried out on splenocytes collected from athymic nude mice. As shown in Figure 3A
, splenocytes from nude mice failed to proliferate with Gal-8 or Con A, although as expected, responded well to LPS. On the other hand, addition of 2 µM Gal-8 led to cpm values twofold higher in B cell-depleted suspensions (content of T cells, >97%) than in nondepleted ones (Fig. 3B)
. Taken together, these results strongly support T cells as the principal targets for Gal-8-induced proliferation. To explore these observations further, purified CD4+ and CD8+ lymphocytes were tested separately in proliferation assays in response to the Gal-8 stimulus. CD4+ cells strongly proliferated; meanwhile, the CD8+ population was comparatively unresponsive (Fig. 3C)
. These findings indicate the strong effect exerted by Gal-8 and support an antigen-independent stimulating effect on CD4+ T cell populations that might be involved in fueling inflammatory processes.
 View larger version (16K): [in a new window] |
Figure 3. Gal-8 induces CD4+ T cell proliferation. (A) Splenocytes from NIHnu/nu mice or their euthymic counterparts were assayed. (B) Splenocytes from C57BL/6J mice depleted of B cells by anti-CD19-coated bead treatment (<2.5% residual B cells) were tested. (C) Gal-8-induced proliferation on purified T cells. Splenocytes from C57BL/6J were subjected to magnetic cell sorting using anti-CD4 or anti-CD8 microbeads (Miltenyi Biotec), following the manufacturer’s instructions for magnetic labeling and separation. Each group (5x105 cells) was incubated with the indicated amounts of Gal-8. (A and B) cpm corresponding to Gal-8 plus TDG were discounted. Results are representative of three to five independent assays. LPS and Con A were added at 5 µg/ml as controls.
|
, and IL-2 and IL-2 was observed in Gal-8-treated cells (Fig. 4
). On the other hand, IL-10, IL-5, IL-13, which were undetectable under the conditions of this assay, and IL-12 showed no significant variations in their transcription rates upon Gal-8 treatment. Identical results were obtained for every cytokine tested when splenocytes were incubated in the presence of Gal-8 for a shorter period of 4 h (data not shown). As IFN- is characteristic of the Th1 response and IL-4 of the Th2 response, we concluded that Gal-8 is probably not associated with the activation of either subset. Although other cells such as NKT are also producers of these cytokines, their numbers (1% of total T cells) are relatively low, so these results, based on mRNA quantitation, cannot be ascribed to them [21
]. These results are in agreement with recent findings obtained with the plant lectin Jacalin, which induces T cell activation through CD45 with a similar cytokine profile [22
], thus suggesting that this lectin might be partially mimicking the endogenous Gal-8 reactivity.
 View larger version (81K): [in a new window] |
Figure 4. Gal-8 induces the expression of IL-2, IFN-, and IL-4. Splenocytes were cultured with Gal-8 alone or in combination with TDG, and a RT-MPCR was performed to determine transcripts for several ILs and GAPDH as control. The position of control bands is indicated on the left.
|
 View larger version (38K): [in a new window] |
Figure 5. Identification of putative Gal-8 counter-receptors by MALDI-TOF. Splenocyte extracts were passed through Gal-8-Sepharose affinity columns and eluted with 100 mM lactose in PBS. Eluted material was solved by 10% SDS-PAGE and numbered bands subjected to MALDI-TOF. For description of all identified molecules, see Table 1
.
|
|
View this table: [in a new window] |
Table 1. Gal-8 Protein Binders Identified by MALDI-TOF from Splenocytes
|
 View larger version (18K): [in a new window] |
Figure 6. Involvement of CD45 in the induction of cell proliferation by Gal-8. (A) Inhibition of Gal-8-induced splenocyte proliferation. Splenocytes were incubated with the indicated amounts of Gal-8. TDG (30 mM) and CD45 PTPase inhibitor (Inh; 0.5 and 2 µM) were tested. *, P < 0.05; **, P < 0.01. Control cpm values corresponding to Gal-8 plus TDG were discounted in all groups. No significant differences were found with basal cpm in cultures where only TDG or PTPase inhibitor was added. (B) Gal-8-induced pERK1/2 in Jurkat T cells. (C) Inhibition of pERK prevents splenocyte proliferation. The U0126 inhibitor was added at 10 µM 1 h before Gal-8. LPS and Con A were used at 2.5 µg/ml and included as controls. LPS- or Con A-induced proliferation alone was not inhibited by the CD45 PTPase inhibitor addition (not shown). cpm values corresponding to U0126 alone were similar to basal cpm. *, P < 0.0002 ; **, P < 0.0004. Assays were performed at least three times.
|
 View larger version (27K): [in a new window] |
Figure 7. Costimulatory activity of Gal-8. (A) Splenocytes (1.5x105 nucleated cells) from DO11.10 mice were cultured for 48 h in the presence of the cognate OVA323–339 peptide (OVA; 0.4–1 µg/ml) together with a low amount of Gal-8 (0.1 µM). Assays were performed at least three times. *, P < 0.01. (B) Same as before, but higher concentrations of Gal-8 were tested and cultures collected at 24 h. When Gal-8 was assayed at 0.5 or 1 µM, the OVA (black bars) versus OVA + inhibitor (dark gray bars) groups: P < 0.05. No significant differences were found between Gal-8 2 µM groups. (C) pZAP70. The experiment was performed as previously, but 5 x 106 cells were incubated for 1 h. Cells were collected and processed for Western blotting. (D) Splenocytes from DO11.10 mice were preincubated with 10 µM of the ERK inhibitor U0126 for 1 h before the addition of Gal-8 and/or OVA. Experimental conditions are the same as in A. *, P < 0.0002.
|
The ability of Gal to manipulate the immune system is receiving strong interest. There is supportive evidence regarding Gal-1 and its capacity of promoting apoptosis of T cells during their development in the thymus as well as in the periphery after activation [2 , 30 31 32 33 ]. Gal-2, structurally related to Gal-1, also promotes apoptosis of activated T cells [34 ]. Gal-3, the only chimera-type Gal, promotes T cell apoptosis when exogenously added [35 , 36 ], and Gal-9, which belongs to the same group as Gal-8, has been shown to induce apoptosis of immature thymocytes and of activated peripheral Th1 cells [4 , 37 ]. In another work, Gal-9 has been shown to induce apoptotic death in Jurkat T cells in contrast with Gal-8, which showed no apoptosis [38 ]. This is in agreement with the proliferative response and the absence of apoptosis induced by Gal-8 in peripheral lymphocytes reported here. These data highlight further the differences observed in the induced cell behavior among the different Gal, which in spite of their similar sugar recognition [39 ], may induce different and even contrasting effects on the immune cells.
The use of different lectin concentrations and culture periods allowed us to discriminate between two different biological activities for Gal-8 on peripheral T cells. The addition of Gal-8 to cultures induced strong antigen-independent proliferation on T cells (Figs. 2
, 3
, and 6
) and a costimulatory effect on a given T cell response (Fig. 7)
. Although all populations in the spleen were targeted by Gal-8 (Fig. 2A)
, the proliferative effect could be associated only with CD4+ T cells by cell depletion and assays with athymic mice. No proliferation was observed in B cells, although the CD45R isoform was found as a counter-receptor of Gal-8 by MALDI assays performed with splenocytes from euthymic (Fig. 5
and Table 1
) and athymic mice (not shown). CD4+ T cells were stimulated by the Gal, and the increased expression of IL-2, IFN-, and IL-4 suggests that Th1 and Th2 populations were induced to proliferate (Fig. 4)
. Several integrins are targets of Gal-8 (see refs. [13
14
15
16
] and Table 1
) and induce biological responses such as adhesion and spreading in Jurkat and other cells. Therefore, they may also be potentially responsible for the proliferative outcome reported here. In fact, stimulation of LFA-1, identified here as a Gal-8 counter-receptor, induces IL-2 production in Jurkat cells [40
]. CD45 was also identified as a Gal-8 counter-receptor by MALDI-TOF analysis, and inhibition of its PTPase activity abrogated Gal-8-induced proliferation almost completely. These results support that Gal-8 acts as a T cell-activating molecule by agonistic binding to CD45. Perhaps it is associated with the effect exerted by CD45 reported recently, which lowers the T cell activation threshold [41
, 42
]. In contrast, anti-CD45 antibodies did not prevent the apoptosis induced by Gal-9 [38
]. These data denote a biological function for Gal-8 on mature T cells that distinguishes it from other Gal studied until now. Jacalin, a plant lectin that binds to CD45, induces the secretion of the same pattern of ILs and pERK1/2 as Gal-8 [22
], suggesting that this lectin might be mimicking the endogenous Gal-8 partially.
Gal-8 was able to synergize the TCR-specific signaling, thus enhancing a specific T cell response induced in the presence of a low dose of the cognate peptide (Fig. 7A) . This ability was disclosed when the lectin was assayed at such low doses that it was unable to induce the polyclonal proliferation. As the lectin dose increases, the differential effects between costimulation and polyclonal proliferation become overlapped as a result of the strong proliferation induced by Gal-8 in the absence of antigen (Fig. 2) . However, the costimulatory effect was observed readily when cultures were collected earlier at 24 h of stimulation instead of at 48 h (Fig. 7 A vs. B) . In these assays, the antigen-independent stimulation showed a delay in the onset, which allowed us to discriminate both activities, indicating that other activation pathways different from the fast TCR signaling are involved in that process. When high Gal-8 doses such as 2 µM were tested, no costimulatory activity was observed (Fig. 7B) ; this might be ascribed to the observed agglutination of the cells that precluded the proper antigen presentation. This absence of costimulation is in fact supporting the activation of independent but synergizing signaling pathways, as at this high dose, the proliferative capacity of Gal-8 was observed readily (Fig. 2) . In agreement, CD45 PTPase inhibition prevented the proliferation only partially (not shown), and pERK1/2 was fully abrogated by TDG but reduced only partially by CD45 PTPase inhibition in Jurkat cells treated with Gal-8 (Fig. 6) . The costimulatory activity was observed at lower Gal-8 concentrations (Fig. 7A) , thus suggesting that might be the primary in vivo activity of this lectin. This costimulatory effect was prevented by the addition of the CD45 PTPase inhibitor, probably as Gal-8 triggered the TCR pathway by working as an endogenous agonist for the extracellular domain of CD45 [41 , 42 ]. In agreement, pZAP70 was demonstrated readily with Gal-8 at levels similar to those observed with the cognate peptide stimulus (Fig. 7C) . These findings allow us to postulate that in vivo, this lectin might help trigger some otherwise borderline signals. Activities reported here help to understand the opposite outcome observed on thymocytes, where even when binding to all populations, Gal-8 induces apoptotic death only in the CD4highCD8high cells [7 ]. The binding of the TCR with high affinity is an event that in the thymus, most probably leads to apoptosis, whereas in the periphery, leads to cell activation. This might be mimicked by the activation of the CD45 PTPase, which is known to lower the TCR signaling threshold and to exacerbate the response [41 , 42 ]. Therefore, the external stimulation of this molecule by Gal-8 will urge T cells to proliferate and to induce the double-positive thymocyte apoptosis reported previously [7 ]. In support, Gal-1 has been involved recently in agonistically signaling in thymic selection and in peripheral activation of CD8+ T cells [43 ].
Gal-8 is expressed abundantly in many tissues [13 ], including the injured endothelium [44 ], and has been shown recently to reach high concentrations (estimated at 25–65 nM) in the inflamed sinovia [45 ], which are close to those used here in vitro. Although Gal-8 was postulated to exert an apoptotic effect on inflammatory sinovial cells, this is based on the increase in Annexin-V binding [45 ], an event described more recently as not necessarily associated with cell death induction [38 , 46 , 47 ]. Interestingly, the exposure of phosphatidylserine and Ca2+ uptake in Jurkat T cells is observed at high concentrations of Gal-8 but in the absence of DNA fragmentation [38 ]. Instead, exposure of phosphatidylserine is related to cell activation by this lectin [47 ], thus supporting the proliferative activity reported here as probably involved in inflammatory outcomes. In agreement with this, Gal-4 is able to stimulate CD4+ T cells to activate and exacerbate inflammatory bowel disease [48 ]. In addition, several markers from inflammatory cells such as neutrophils (Mac-1 and MMP-9 reported previously as counter-receptors for Gal-8 [17 ] and the CD177 antigen) as well as macrophages and platelets were also identified as Gal-8 counter-receptors (Table 1) . Therefore, Gal-8 seems to be involved in the migration or activation of these inflammatory cells, which together with the T cell-proliferative activity reported here, supports a fueling role in the pathogenesis of inflammatory diseases. Importantly, Gal-8 was also able to costimulate T cells, even at low concentrations, an activity that might be functional, not only under normal conditions but also in altered responses such as in the induction of autoimmunity where the lowering of the threshold of T cell activation seems crucial.
ACKNOWLEDGMENTS
This work was supported by grants from the Agencia Nacional de Promoción de la Ciencia y la Tecnología (ANPCyT), Universidad Nacional de San Martín, and Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET; Argentina). M. V. T. is a Fellow, and J. M. and O. C. are researchers from CONICET. V. C. is a Fellow from ANPCyT. We thank Dr. Yehiel Zick (The Weizmann Institute of Science) for providing us with an aliquot of the mouse anti-rat Gal-8 mAb and Dr. Anne I. Sperling (University of Chicago) for providing us with the B6.CD43null mice. We are indebted to Dr. Carlos Buscaglia for his critical reading of the manuscript. The technical assistance of Mr. Fabio Fraga in animal care is appreciated.
FOOTNOTES
Abbreviations: Gal-8=Galectin-8, Mac-1=macrophage antigen-1, MMP=matrix metallopeptidase, MPCR=multiplex PCR, NIHnu/nu=NIH nude, p=phospho, PTPase=phospho-tyrosine phosphatase, TDG=thiodigalactoside
Received September 5, 2008; revised February 23, 2009; accepted March 23, 2009.
REFERENCES
M Glycobiology 13,755-763
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
