The thiol redox state of lymphoid organs is modified by immunization: Role of different immune cell populations^†

Lymphoid tissues are embedded with large amount of non-protein thiols (NPSH)

Tissue extract analyses revealed that the thiol content of spleen from germ-free mice is slightly higher than that of non-lymphoid organs, such as kidney and liver, and increases dramatically after immunization (Fig. 1A). To gain information about the intra-tissue distribution of thiols, frozen tissue sections from non-lymphoid and secondary lymphoid organs were stained with mercury orange, a dye specific for non-protein thiols (NPSH) 9-11 and analyzed by confocal microscopy. Liver and kidney displayed a diffuse but faint staining (Fig. 1B, panels 1f and 1g), while LN and spleen from non-injected mice were mildly positive (Fig. 1B, panels a, c and h). After Ag injection, both LN and spleen increased the positivity to NPSH staining (Fig. 1B, panels b, d and i). The mercury orange signal did not appear restricted inside cells but was also present in the intercellular space, suggesting that NPSH are externalized. Pretreatment of cryostat sections with the thiol-specific blocking agent n-ethylmaleimide completely abrogated the mercury orange staining (Fig. 1B, panels e and m), confirming the thiol-selectivity of the dye 10, 12. Following immunization, the MFI of inguinal LN increased from 1.5- to 2-fold and that of mesenteric LN up to 10-fold (Fig. 1C). Portal LN resulted 2-fold more intensely stained by mercury orange in mice with a Con A-induced liver injury 13 than in untreated mice, indicating that NPSH are induced also in intra-visceral LN upon activation (Fig. 1C). In contrast, no fluorescence increase was observed in the injured liver (not shown). Together, these results indicate that lymphoid tissues are significantly more “reduced” than the other tissues, and that the immune response is paralleled by a conspicuous increase in the NPSH production. GSH is the major antioxidant NPSH inside cells, but irrelevant outside. At variance, even if usually in very low concentrations in the extracellular space, the Cys/CySS pool is the major low-molecular-weight thiol/disulfide couple in plasma, and serves important antioxidant functions 14. Therefore, although mercury orange staining does not discriminate between GSH and Cys, it is conceivable that the intracellular staining is due mainly to GSH, while the extracellular to Cys.

Gut-associated Peyer's patches were the only lymphoid organs almost negative to mercury orange staining (Fig. 1C, GALT) and non-activated mesenteric LN were significantly (p<0.001) less stained than inguinal LN. This feature may be related to the physiological unresponsiveness to antigen exposure in the intestinal microenvironment 15. In this regard, lamina propria macrophages have been found unable to release Cys, and thereby responsible for the low proliferative potential of lamina propria T lymphocytes 16. Moreover, mesenteric LN have been identified as the key site for tolerance induction to food antigens, mediated by DC 17. The dramatic increase of NPSH in mesenteric LN after Ag injection indicates that this deficit is not permanent but is overcome by a powerful immunization, such as i.p. injection. This suggests that a dysregulation of GALT redox may be involved in inflammatory bowel disease 18, 19. However, it cannot be ruled out that in non-immunized mice the higher staining of superficial LN with respect to LN associated with the gastrointestinal tract is due in part to a higher basal activation, related to an easier accessibility of superficially located LN to external noxia.

NPSH colocalize with DC and increase in B-cell areas of activated LN

Double-staining with mercury orange and B, T or DC markers (Fig. 2) showed that NPSH staining of non-immunized LN (panel A) was generally diffuse in the cortex and paracortex zones with no clear compartmentalization between B zone (upper left panels, CD45R⁺) and T zone (middle left panels, CD3⁺). However, a stronger staining was found to cluster with DC distributed in the T zones as indicated by the co-localization of CD205 and mercury orange highly positive cells (lower left panels). This observation suggests that, like in vitro 5, also in vivo during Ag presentation DC produce Cys in the extracellular space, making it available for T-cell activation and proliferation. In immunized LN (Fig. 2B), the dramatic increase in NPSH staining was highly intense in the B zone of the outer cortex, especially in the GC. This finding suggests that a reducing environment is needed by B cells activation in vivo, confirming the long standing evidence that B lymphocyte cultures require reductants such as 2-ME 4. Mercury orange staining condensed in highly fluorescent large spots within the GC. Hematoxylin/eosin staining revealed the presence of abundant nuclear and cell debris (Fig. 2C, upper panel), indicating that these spots contain remains from dying B cells. This was further confirmed by the presence of MFG-E8⁺ TBM in the duty to engulf apoptotic B cell (Fig. 2C, lower panel) 20. The GC spots containing cell debris and highly stained by mercury orange are surrounded by T lymphocytes (Fig. 2B middle panels, arrowheads) that may be engaged in interactions with blebs from apoptotic B cells, a process that has been proposed to limit the availability of T-cell help 7.

B cells and macrophages release NPSH after activation

The strong NPSH content in GC of activated LN suggested the involvement of activated B cells in NPSH generation. Indeed, while resting B cells isolated from human tonsils B lymphocytes by Percoll gradient (high-density B cells) 21 did not release detectable NPSH over 12 h of culture, in vivo activated B lymphocytes (low-density B cells) released levels of NPSH varying from 10 to 20 μM in the different subjects (Fig. 3A). A similar amount of released thiols was detected in supernatants from high-density B cells after 12 h of exposure to the B cell mitogen Staphylococcus aureus Cowan I (SAC). That activated B lymphocytes secrete Cys is coherent with their APC function 22 and indicate that they can provide T cells with exogenous Cys during antigen presentation. On the other hand, recent evidence indicates that B cells undergo oxidative stress during differentiation 23 and elicit an adaptive response with increased expression of peroxiredoxins and other anti-antioxidant genes. Therefore, it is possible that the large amounts of released thiols in GC represent an anti-oxidant response that allows differentiating B lymphocytes to resist otherwise fatal stress. Unlike activated B lymphocytes and DC, both resting and PHA-activated T lymphocytes did not release detectable thiols. In contrast, in vitro derived macrophages were found to secrete the highest levels of NPSH. These data are in agreement with previous findings of high levels of Cys secretion by peripheral blood monocytes 16 but is in contrast with other observations that activated macrophages produce a burst of superoxide anion 24. This apparent discrepancy may be related to an adaptive anti-oxidative response, with increased expression of Trx and NPSH, triggered by the oxidative stress as observed in different cell systems 11, 23. The high level of NPSH released by activated monocytes in vitro supports the observation of the strong positivity in the LN areas occupied by TBM. The major function of TBM so far documented is the removal of apoptotic B cells, a process implicated in avoiding autoimmunity 20. On these basis, it is tempting to speculate that beside a role in promoting the immune response, the reduced environment may also help macrophages to clear out dead B lymphocytes and thereby to prevent autoimmune diseases.

Trx expression parallels thiol release

Trx is one of the main intracellular oxido-reductases and plays critical roles in the regulation of the cellular redox environment 25. As shown in Fig. 3B, Trx expression, like thiol release, is undetectable in resting but induced in activated B cells. Mature DC and macrophages, that release the highest amounts of NPSH, also contain abundant Trx. Exceptions are PHA-activated T cells that do not release NPSH in spite of expressing more Trx than resting T lymphocytes. The implication of Trx in the generation of NPSH was supported by the inhibition of NPSH release observed after cell exposure to 1-chloro-2,4-dinitrobenzene (DCNB), a compound that blocks Trx function 5 (Fig. 3C). While non-immunized spleen and LN do not stain detectably with anti-Trx (not shown), lymphoid tissues from immunized mice (Fig. 3D) showed a diffuse positivity for Trx, more intense in macrophage areas around B-cell areas (see for comparison the positivity for MOMA-1⁺ macrophages in the right panel) and in GC.

T and B Lymphocytes modulate the expression of the cys-SH transporter

The cystine/glutamate antiporter x is the main transport system that mediates the cellular uptake of Cys in exchange for intracellular glutamate 26. As shown in Fig. 3C, glutamate inhibits the generation of extracellular NPSH by macrophages and DC, suggesting that CySS is indeed internalized through x. Accordingly, DC and monocytes expressed the x light chain xCT (Fig. 3E). At variance, neither resting B lymphocytes nor resting T cells express xCT. However, a short (3 h) stimulation with SAC or PHA of B and T cell, respectively, induces the expression of xCT (Fig. 3E). Therefore, both resting B and T cells require exogenous Cys for their activation, but this requirement is essential only during the first phases of the immune response, and it is rapidly overcome by the expression of x once cells are activated.

Chemicals and animals

All chemicals were from Sigma (Milano, Italy) unless differently specified.

BALB/c mice from Charles River (Lecco, Italy) were maintained in pathogen-free facilities. Twelve-week-old animals (three per group) were injected weekly twice with 100 μL of a 1:1 BSA (1 mg/mL)/CFA emulsion ratio, s.c. or i.p. to stimulate superficial and mesenteric LN, respectively. Liver toxicity and portal LN activation were induced by i.v. injection of 400 μg of Con A 13. Spleens were immunized by a single i.v. injection or four times weekly i.p. injections of 200×10⁶ SRBC. One week after the last injection, mice were euthanized and the different organs were removed and analyzed. All protocols were approved by the Institutional Review Board of the Istituto Nazionale per la Ricerca sul Cancro (Genova, Italy).

Thiol measurement

Liver, kidney, spleen and LN were homogenized and processed as described 29. Thiols in clarified homogenates and in cell culture supernatants were quantified using the 5,5′-dithiobis(2-nitrobenzoic acid) (DTNB) method as described 5, 29. L-Cys was used as standard. Mean and SD were calculated using GraphPad Software. For NPSH quantification in supernatants of cells treated with DCNB or glutamate, standard curves with L-Cys were supplemented with DCNB or glutamate at the same doses used in culture. No significant differences were detected between the standard curves without or with the two compounds.

Immunohistochemical and immunofluorescence procedures

Serial cryostat sections from frozen organs were reacted with 25 μM mercury orange 9-11 and co-stained with the following Ab: rat anti-mouse mAb to CD45R/B220 (B cell specific), CD205 (DC specific), CD3ε (T-cell specific) from Southern Biotech (Birmingham, AL) and rabbit-anti-mouse MFG-E8 Ab (TBM specific) from DBA (Milan, Italy) followed by the relevant anti-IgG-FITC-conjugates (Southern Biotech), and analyzed by confocal microscopy 11. FITC and mercury orange were excited with 488 and 543 nm laser, respectively. The emission signal was filtered at 505–530 and 540 nm. Excitation lasers were used successively, to avoid overlapping signals. Images were acquired and quantified by the Fluoview FV500 software. When indicated, sections were pretreated 10 min with 100 μM n-ethylmaleimide before mercury orange staining 12.

For immunohistochemistry, mouse mAb to Trx (2B1, kind gift of Dr. F. Clarke, Brisbane, Australia) and rat mAb to mouse MOMA-1 (MΦ specific, Southern Biotech) were used. Immunostaining was performed using a streptavidin–biotin–alkaline-phosphatase-complex staining kit 11 (Bio-SPA Division, Milan, Italy). For each group of organs three representative zones have been sampled.

Cell preparation and cell cultures

T cells were isolated from PBMC from healthy donors as described 5 and stimulated 3 or 15 h with 10 μg/mL of PHA-P as indicated. Macrophages and DC were obtained by culturing adherent cells from PBMC as described 5.

Tonsils were obtained from children undergoing tonsillectomies, after informed consent. B cells were purified and fractionated by Percoll density gradients as described 20 and activated 3 or 15 h with SAC (Calbiochem, Milano, Italy).

For NPSH-release assay, a different cell number was plated according to the cell dimension (DC: 0.25×10⁶, MΦ: 0.4×10⁶, B and T lymphocytes: 4×10⁶).

Western blot analysis

Western blot analysis was performed on cell lysates as described 5. Filters were hybridized with anti-Trx mAb 2B1 or anti-actin mAb (Sigma).

RT-PCR

Total RNA was isolated from 1×10⁷ cells using TriPure Isolation Reagent (Roche, Indianapolis, USA). For RT-PCR, the first-strand of cDNA was synthesized from 1-μg of the total RNA using random examers and M-MuLV RT (Roche). PCR amplification of cDNA was performed using the primer set: 5′-ATG GTC AGA AAG CCT GTT G-3′/5′-TCT TCT GGT ACA ACT TCC AGT ATT A-3′, designed from the nucleotide sequences of human xCT 30 and the primer set 5′-CCT GCT TGC TGA TCC A-3′/5′-TCC GGT GAC GGG GTC A-3′ to amplify human β-actin gene, as internal control.