B. Rearte, Academia Nacional de Medicina, Pacheco de Melo 3081, C1425AUM Buenos Aires, Argentina. E-mail: firstname.lastname@example.org
Sepsis and septic shock can be caused by Gram-positive and -negative bacteria and other microorganisms. In the case of Gram-negative bacteria, endotoxin, a normal constituent of the bacterial wall, also known as lipopolysaccharide (LPS), has been considered as one of the principal agents causing the undesirable effects in this critical illness. The response to LPS involves a rapid secretion of proinflammatory cytokines such as tumour necrosis factor (TNF)-α, interleukin (IL)-1, IL-6, interferon (IFN)-γ and the concomitant induction of anti-inflammatory mediators such as IL-10, transforming growth factor (TGF)-β or glucocorticoids, which render the host temporarily refractory to subsequent lethal doses of LPS challenge in a process known as LPS or endotoxin tolerance. Although protective from the development of sepsis or systemic inflammation, endotoxin tolerance has also been pointed out as the main cause of the non-specific humoral and cellular immunosuppression described in these patients. In this report we demonstrate, using a mouse model, that mifepristone (RU486), a known glucocorticoid receptor antagonist, could play an important role in the restoration of both adaptive humoral and cellular immune response in LPS immunosuppressed mice, suggesting the involvement of endogenous glucocorticoids in this phenomenon. On the other hand, using cyclophosphamide and gemcitabine, we demonstrated that regulatory/suppressor CD4+CD25+forkhead boxP3+ and GR-1+CD11b+ cells do not play a major role in the establishment or the maintenance of endotoxin tolerance, a central mechanism for inducing an immunosuppression state.
The development of new therapies for sepsis has been particularly frustrating during the last 30 years, when more than 25 trials of new agents have failed  in a period referred to as the ‘graveyard for pharmaceutical companies’.
This failure has been due partly to a lack of understanding of the pathogenic mechanisms driving sepsis . In fact, most therapies for sepsis have focused upon attenuating the initial inflammatory response, ignoring the progressive development of immunosuppression developed in the anti-inflammatory phase [4,5], which is the cause of the inability to fight secondary infections in the post-septic period .
Although these approaches have demonstrated modest benefits in selected groups of patients, the majority of deaths occur in septic patients who are immunosuppressed [5,7–11]. In addition, it has been reported recently that dormant viruses such as herpes simplex and cytomegalovirus can also be reactivated in these patients [12,13].
In sepsis and septic shock caused by Gram-negative bacteria, endotoxins, a normal constituent of the bacterial wall also known as lipopolysaccharide (LPS), has been described as one of the principal agent causing the undesirable effects in this critical illness .
The response to LPS involves a rapid secretion of proinflammatory cytokines [tumour necrosis factor (TNF)-α, interferon (IFN)-γ, interleukin (IL)-1, IL-6, IL-8] and the concomitant induction of anti-inflammatory mediators such as IL-10 and transforming growth factor (TGF)-β or glucocorticoids [2,14–18], which render the host temporarily refractory to subsequent lethal doses of LPS challenge in a process known as LPS or endotoxin tolerance [19–22].
Although protective from the development of sepsis or systemic inflammation, endotoxin tolerance has also been pointed out as the main cause of the non-specific humoral and cellular immunosuppression described in these patients [12,21,23,24]. Taking this into account, and in agreement with Hotchkiss et al. , we think that one of the key aspects of late sepsis is immunosuppression, and that re-engaging or preserving host immune functions will be the next major advance in the management of this illness.
Recently, we obtained preliminary results about the restoration of the primary humoral immune response through the use of mifepristone (RU486), an inhibitor of glucocorticoid receptors , in a model of LPS-induced tolerant/immunosuppressed mice.
In order to gain insight into this effect, we extended our experimental work aiming to restore both secondary humoral and cellular immune responses. For the latter we used a tumour mouse model, where adaptive T cell-mediated immune response is crucial to avoid the tumour takes.
In addition, we analysed the eventual importance of immature myeloid-derived suppressor (GR-1+CD11b+) and regulatory T [CD4+CD25+ forkhead box P3 (FoxP3)+] cells, two types of cells claimed to be involved in different mechanisms of immunosuppression, including those induced by sepsis [26–28]. For this purpose we used cyclophosphamide, a cytostatic agent that induces a depletion of regulatory T cells , and gemcitabine, which deletes myeloid-derived suppressor cells in mice [26,30].
In brief, our results indicated that RU486 could play an important role in the restoration of both adaptive humoral and cellular immune response in LPS immunosuppressed mice, suggesting the involvement of endogenous glucocorticoids in this phenomenon. Conversely, using cyclophosphamide and gemcitabine, we demonstrated that regulatory/suppressor CD4+CD25+FoxP3+ and GR-1+CD11b+ cells would not play a major role in the establishment or maintenance of endotoxin tolerance, a central mechanism for inducing an immunosuppression state.
Materials and methods
Mifepristone [RU486-17-hydroxy-11-(4-dimethylaminophenyl) 17-(1-propynyl) estra-4,9-diene-3-one], LPS from Escherichia coli O111:B4 purified by phenol extraction and propyleneglycol (1,2-propanediol), were obtained from Sigma-Aldrich (St Louis, MO, USA). Dexamethasone (Decadrón Shock) was obtained from Sidus S.A. (Buenos Aires, Argentina). Cyclophosphamide was obtained from Lab Kampel Martian (Buenos Aires, Argentina). Gemcitabine was obtained from Gemtro® (Eli Lilly and Company, Indianápolis, IN, USA). Reagents were prepared in sterile pyrogen-free saline, except RU486, which was dissolved in propyleneglycol. Sheep red blood cells (SRBC) were obtained from Alfredo Gutiérrez® (Buenos Aires, Argentina).
BALB/c mice were bred in the animal facility of the Academia Nacional de Medicina, Buenos Aires. Mice aged between 12 and 16 weeks weighing 20–25 g were used throughout the experiments. They were maintained under a 12-h light–dark cycle at 22 ± 2°C and fed with standard diet and water ad libitum. The experiments performed herein were conducted according to the principles set forth in the Guide for the Care and Use of Laboratory Animals.
Endotoxin tolerance models
Classical tolerance model. Mice were tolerized by intraperitoneal (i.p.) inoculation of LPS (5 µg/mouse) for 3 consecutive days. Twenty-four h after the last injection animals were resistant to a lethal dose (LD) of LPS (2 LD50 = 200 µg/mouse; i.p.).
Tolerance/immunosuppression model. Because immunosuppression is a quantitative effect dependent upon the number of doses and concentration of LPS injections, stronger immunosuppression was obtained by treatment of mice with different doses of LPS for 12 consecutive days. The inoculation regimen began with 5 µg/mouse i.p. for the first 4 days, followed by 100 µg/mouse i.p. for 8 days.
Mice were injected i.p. with a lethal dose of LPS (2 LD50 = 200 µg/mouse) in pyrogen-free saline and followed up to 72 h. This dose induces 100% mortality between 24 and 72 h after injection. The same batch of LPS was used throughout the experiments.
Secondary humoral immune response and RU486 treatment of tolerant/immunosuppressed mice
Mice were first immunized with a cell suspension of sheep red blood in phosphate-buffered saline (PBS) (SRBC; 5 × 108 cells/mouse; i.p.). The immunosuppression scheme with LPS began 15 days after immunization by daily i.p. injections of 5 µg of LPS/mouse/day during 4 days, followed by 100 µg LPS/mouse/day during the last 8 days. Then, 24 h after the last dose of LPS, mice were treated, or not, with RU486 (30 mg/kg; i.p.) and 30 min later they were immunized with a second dose of SRBC (5 × 108 cells/mouse; i.p.). Twenty-four and 30 h after immunization, mice were treated, or not, again with RU486. Control mice were inoculated with SRBC using the same schedule. Seven days after the second immunization the animals were bled and the sera were collected to evaluate the secondary humoral immune response.
Methylcholanthrene-induced murine fibrosarcoma (MC-C) tumour and X-irradiated tumour cells
MC-C is a methylcholanthrene-induced fibrosarcoma obtained in 5-month-old BALB/c mice. Tumour was used between subcutaneous (s.c.) passages 5–25. Non-necrotic tumour sections were minced, passed through a sterile mesh and a single cell suspension was obtained. Then, cells were irradiated to immunize mice as described previously .
Cellular immune response and RU486 treatment of tolerant/immunosuppressed mice
LPS immunosuppressed mice were treated, or not, with RU486 (30 mg/kg; i.p.) and 30 min later were inoculated with irradiated MC-C tumour cells (immunization; 6 × 106 cells/mouse; s.c.). Twenty-four and 30 h after immunization, mice were treated, or not, again with RU486. Seven days after immunization, mice were challenged with live MC-C tumour cells (2 × 105 cells/mouse; s.c.) and treated, or not, with RU486 using the same schedule defined for the immunization step. Naive mice were immunized and challenged using the same schedule. Control mice were challenged with live MC-C tumour cells. Ten days after tumour challenge we evaluated the tumour growth considering presence or absence of tumour.
Antibody response to SRBC was evaluated through a haemagglutination assay, as described previously . Briefly, serum samples were diluted at an initial 1:50 dilution and 200 µl of the dilution was added to the first well and was serially diluted (1:0·25). The titre was considered as the reciprocal of the last positive dilution.
Drug treatment of tolerant mice
LPS-tolerant/immunosuppressed mice were treated with cyclophosphamide (150 mg/kg, i.p.), gemcitabine (120 mg/kg, i.p.) or saline 24 h after the last LPS dose. Control mice (naive) were treated, or not, with the same schedule. After 48 h, depending on the experiment, mice were either killed to harvest splenocytes or challenged with a lethal dose of LPS.
Drug treatment during the establishment of endotoxin tolerance
Mice were injected with cyclophosphamide (100 mg/kg; i.p.), gemcitabine (120 mg/kg; i.p.) or saline. After 48 h the animals were inoculated with a LPS tolerizing dose (5 µg/mouse; i.p.) concomitantly to an additional dose of the same drug. Control mice (naive) were treated, or not, with the same schedule of drugs. The protocol included two additional doses of LPS and 48 h later mice were either euthanized to harvest splenocytes or challenged with a lethal dose of LPS.
Isolation of splenocytes
Spleens were removed and passed through sterile mesh to obtain a single cell suspension. Erythrocytes were eliminated with lysing buffer and cells were washed and analysed by flow cytometry.
To measure immunoglobulin G (IgG) anti-SRBC, mice sera were diluted in PBS-bovine serum albumin (BSA) 0·5% and incubated with SRBC 1% for 30 min at 4°C, as described previously . Then, SRBC were washed and incubated with fluorescein isothiocyanate (FITC)-labelled anti-IgG antibodies (Jackson Immuno-Research Laboratories, West Grove, PA, USA) for 30 min at 4°C. Finally, cells were washed and immunoglobulins evaluated.
Splenocytes were stained with a combination of the following monoclonal antibodies (mAbs): FITC-CD11b, phycoerythin (PE)-Ly6G (GR-1) (BD Pharmingen, San Diego, CA, USA) and PE/Cy5-F4/80 (BioLegend, San Diego, CA, USA). The latter was used to exclude macrophage populations. To evaluate regulatory T cells we used FITC-CD4, peridinin chlorophyll (PerCP)-CD25 (BD Pharmingen) and PE-FoxP3 (eBioscience, San Diego, CA, USA). Cells were suspended in 10 µl of PBS-1% normal mouse serum (15 min) and incubated with the corresponding antibodies for 30 min at 4°C, washed, fixed in PBS/0·5% paraformaldehyde and evaluated in a Becton Dickinson fluorescence activated cell sorter (FACScan) using FSC Express software (De Novo Software, Los Angeles, CA, USA). For FoxP3 staining, cells were stained intracytoplasmatically using a permeabilization/fixations set according to the manufacturer's instructions (FoxP3 Staining Buffer Set; eBiosciences).
Values are expressed as the mean ± standard error of the mean (s.e.m.) of n observations. Differences among groups in flow cytometric analyses were evaluated by one-way analysis of variance followed by Tukey's multiple comparison test. Haemagglutination assays were analysed using Student's unpaired t-test. Differences in tumour frequency were determined by χ2 test. All statistical tests were interpreted in a two-tailed fashion and P < 0·05 was considered significant.
RU486 restores the secondary humoral immune response in LPS immunosuppressed mice
Endotoxin tolerance has been considered to be the initial phase and one of the causes of the immunosuppression observed frequently in late sepsis [20,23,33,34]. We have demonstrated recently that RU486 was capable of inducing a transient overcoming of LPS-induced tolerance/immunosuppression, restoring primary humoral immune response in immunosuppressed mice, either at the level of IgM or IgG antibodies .
In this paper we extended our studies to the secondary humoral immune response. For this purpose, three groups of mice (n = 9/group) were injected with sheep red blood cells (SRBC) in order to mount a normal primary response. Then, 15 days later, animals from two of the three groups were immunosuppressed with LPS. Then, 24 h after the last injection, one of the two LPS-treated groups was treated with RU486 and all groups were injected again with SRBC. Seven days later the secondary humoral immune response was analysed. The results shown in Fig. 1a and b indicate clearly that LPS induces profound immunosuppression and that RU486 can, partially although significantly, restore the secondary humoral immune response to SRBC in LPS immunosuppressed mice.
RU486 restores the cellular immune response in LPS immunosuppressed mice
In order to determine if RU486 was capable of overcoming the LPS-induced immunosuppression in a T cell primary response, we used a strongly immunogenic MC-C, which can be rejected in preimmunized mice through T cell-dependent mechanisms .
For this purpose, two groups of previously LPS immunosuppressed BALB/c mice were immunized with irradiated MC-C tumour cells in the presence or absence of RU486. Then, 7 days after immunization, MC-C cells were implanted subcutaneously, again in the presence or absence of RU486, and tumour progression of the two groups were compared to a standard implantation in naive mice (control).
As can be seen in Table 1 and in accordance with previous results , MC-C tumour grew in 100% of naive mice while it was inhibited completely in immunized mice. Conversely, MC-C tumour grew in a significant percentage of immunized mice that had been suppressed previously with LPS.
Table 1. RU486 partially restores the cellular immune response in lipopolysaccharide (LPS)-induced immunosuppressed mice.
IS + RU486
P < 0·01 versus IS;
P < 0·001 versus control;
P < 0·001 versus In. LPS-immunosuppressed mice were treated with RU486 or vehicle (propyleneglycol) and immunized with irradiated tumour cells from an immunogenic methylcholanthrene-induced murine fibrosarcoma (MC-C). Naive mice were immunized using the same schedule (In). Seven days later, mice were challenged with live tumour cells and treated, or not, with RU486 using the same protocol defined for the first step. Naive mice were challenged only with live tumour cells (control). Ten days after tumour challenge we evaluated the tumour takes. IS: LPS immunosuppressed group; IS + RU486: RU486-treated immunosuppressed group.
However, RU486 treatment of immunosuppressed mice allowed these animals to inhibit the tumour growth significantly, indicating that RU486 was able to partially restore the primary cellular immune response against the tumour. Finally, RU486 treatment of naive mice at the time of immunization or of the implant of MC-C tumour cells did not modify the normal growth of tumour (data not shown).
Dexamethasone augments GR-1+CD11b+ and CD4+CD25+FoxP3+ cells
Although endogenous glucocorticoids seem to be central molecules in the regulation of LPS-induced immunosuppression, both at the humoral or cellular level, in polymicrobial sepsis models in mice, other authors show an expansion of immature myeloid-derived GR-1+CD11b+ suppressor  and/or CD4+CD25+FoxP3+ regulatory T cells (Tregs)  which have been involved in the immunosuppression state.
Moreover, as it has been demonstrated that glucocorticoids in vitro were capable of inducing a monocyte subset that resembles myeloid-derived suppressor cells , the possibility of involvement of glucocorticoids in the regulation of GR-1+CD11b+ and CD4+CD25+FoxP3+ cells in vivo was explored.
For this purpose, mice were treated with 2·5 mg/kg/day of dexamethasone during 5 consecutive days but, in agreement with previous observations, a dramatic depletion of spleen cells was observed [37,38].
However, considering that the level of corticosterone in tolerant animals was about sevenfold higher than those found in naive animals (tolerants: 1099·6 ηg/ml ± 23·2 versus normal: 163·7 ηg/ml ± 5·8; n = 5) , in order to reach a similar plasma concentration mice were injected with dexamethasone (0·2 mg/kg/day) during 2 consecutive days and spleen cells were analysed.
In this situation a significant increase of GR-1+CD11b+ and CD4+CD25+FoxP3+ cells was observed (Fig. 2a and b), indicating that the appearance of these suppressor/regulatory cells is a consequence of glucocorticoid action.
Role of GR-1+CD11b+ suppressor cells and CD4+CD25+FoxP3+ regulatory T cells in the maintenance of endotoxin tolerance
Taking into account that endotoxin-induced tolerance/immunosuppression is supported by endogenous glucocorticoids (reverted by RU486), that dexamethasone induces the increase of GR-1+CD11b+ and CD4+CD25+FoxP3+ cells and that immunosuppression in tumour and sepsis seems to be the consequence of the expansion of these suppressor/regulatory cells [35,39,40], the presence of GR-1+CD11b+ and CD4+CD25+FoxP3+ cells in the spleen of tolerant mice was analysed.
As shown in Fig. 3a and b, an increase of GR-1+CD11b+ in the spleen of LPS-tolerant mice was observed. We then evaluated the relevance of these cells in the maintenance of endotoxin tolerance. For this purpose, LPS-tolerant animals were treated with gemcitabine, a chemotherapeutic drug that reduces the number of myeloid-derived suppressor cells in mice spleens [26,30].
As shown in Fig. 3a and b, although gemcitabine reduced the LPS-induced GR-1+CD11b+ cells dramatically, maintenance of the tolerance was not modified at all, because animals resist a lethal dose of LPS (200 µg/mouse = 2 LD50), suggesting that these cells were not involved, at least directly, in maintenance of the tolerant state. Similar results were obtained using cyclophosphamide, another cytostatic drug, which induced a strong reduction of these GR-1+CD11b+ cells up to normal levels.
Conversely, tolerant animals have also a slight increase of spleen CD4+CD25+FoxP3+ Tregs, the other cell population involved in immunosuppression.
In order to study the importance of these cells in tolerance, LPS-tolerant mice were then treated with cyclophosphamide (150 mg/kg) and the number of Treg cells was determined. As shown in Fig. 3c, although cyclophosphamide markedly decreased the levels of these regulatory cells from the spleens of tolerant mice, animals resist a lethal dose of LPS, indicating that these cells were not involved in the maintenance of tolerance. Similar results were observed using gemcitabine, although a pronounced effect on Treg cells was not observed.
Role of GR-1+CD11b+ suppressor cells and CD4+CD25+FoxP3+ regulatory T cells in the establishment of endotoxin tolerance
Endotoxin tolerance has two well-defined phases, establishment and maintenance. Establishment is a short period with prevalence of inflammatory cytokines, and maintenance is a long period with predominance of anti-inflammatory agents.
To analyse if GR-1+CD11b+ and CD4+CD25+FoxP3+ suppressor/regulatory cells were involved in the induction or establishment of tolerance to LPS, protocols were designed to eliminate most of these cells during the tolerization period.
Naive mice were then injected i.p. with gemcitabine (120 mg/kg), and 48 h post-injection the animals had a reduced amount of GR-1+CD11b+. Afterwards, an additional dose of gemcitabine was administered concomitantly to the first LPS (5 µg/mouse) injection in order to initiate the establishment of endotoxin tolerance in the absence of GR-1+CD11b+ cells. The protocol included two additional doses of LPS 5 µg. Twenty-four h after the last injection mice were either challenged with a lethal dose of LPS or killed, and spleen cells harvested for flow cytometric analysis. The results depicted in Fig. 4a show that, despite the low levels of GR-1+CD11b+ cells, endotoxin tolerance was established.
In order to elucidate if residual GR-1+CD11b+ cells could eventually have some effect in the establishment of tolerance, we used cyclophosphamide in a treatment that practically eliminated all GR-1+CD11b+ cells. However, all the animals without GR-1+CD11b+ resisted a challenge with a lethal dose of LPS, confirming the results obtained with gemcitabine (Fig. 4a).
Conversely, we observed a slight, although not significant, increase of CD4+CD25+FoxP3+ Tregs during the induction phase of LPS tolerance. Nevertheless, in order to discard any effect of these cells, animals were treated with gemcitabine (120 mg/kg) or cyclophosphamide (100 mg/kg). As shown in Fig. 4b, both drugs induced a significant effect on the reduction of CD4+CD25+FoxP3+ Tregs.
However, here again, despite these treatments all the animals became tolerized. Essentially the same results were obtained by using cyclophosphamide in four different protocols (data not shown)
Taken together, the results depicted in Figs 3 and 4 indicate that neither GR-1+CD11b+ suppressor nor CD4+CD25+FoxP3+ Tregs have any direct influence on the establishment or the maintenance of endotoxin tolerance, as the absence of these cells did not provoke disruption of this phenomenon, the initial phase of immunosuppression.
In recent years the loss of immune competence has been considered to be one of the main problems in late sepsis . In the case of sepsis caused by Gram-negative bacteria, patients show clear signs of endotoxin tolerance [41,42], a phenomenon considered to be the initial phase of immunosuppression [20,23] and thought to play an important role in the susceptibility to reinfection in patients with severe sepsis .
Thus, many investigators consider that efforts to recover or preserve host immune functions will be the next major advance in the management of patients with sepsis .
In agreement with this assumption, in this report we have tried to overcome LPS-induced tolerance/immunosuppression at humoral and cellular levels.
Although glucocorticoids have been used for the treatment of sepsis since as early as 1940, the usefulness of these agents remains controversial and some of their effects are not understood completely [18,21,43]. This lack of understanding may be due to the pleiotropic actions of glucocorticoids [43,44], to the different animal models used, or to conclusions resulting from studies that investigate a particular stage of endotoxin tolerance (i.e. maintenance), later generalized inappropriately.
In a recent study using RU486, an antagonist of glucocorticoids, we succeeded in overcoming LPS-induced tolerance/immunosuppression, restoring the primary humoral immune response , suggesting that corticoids were central agents in this phenomenon. These results encouraged us to evaluate the relevance or influence of corticoids on the secondary humoral, as well as in the T cell-mediated immune responses.
Similar to our previous results obtained in the primary response , RU486 also disrupts LPS-induced immunosuppression and partially restores the secondary humoral immune response. The partial recovery of humoral immune competence indicates that at least some of the cell(s) committed in the secondary response were affected functionally but not killed or eliminated by LPS treatment, and indicate that RU486 could be used to modulate the humoral response during the course of immunosuppression.
The capacity of RU486 to revert immunosuppression was extended to the T cell-mediated immune response. In this case, RU486 restored the capacity to reject MC-C tumours, a property strictly dependent upon specific immune T cells, indicating that glucocorticoids are also one of the most important immunosuppressor agents at the T cell compartment level in LPS immunosuppressed mice.
The reason for the partial recovery of both humoral and cellular immune response could be due to the fact that glucocorticoids induce T cell apoptosis . Therefore, the injection of LPS for 12 days could induce LPS tolerance/immunosuppression accompanied by a decreased T cell number, probably an additional cause of immunosuppression. In addition, cytokines such as IL-10, TGF-β and IL-1Ra have been also mentioned as other possible immunosuppressor agents [5,46].
Despite the fact that we have no evidence of RU486 target cells in the reversion of immunosuppression, because high doses of cyclophospamide did not alter the establishment or the maintenance of tolerance we can speculate that macrophages, key cells in endotoxin tolerance, are possible targets as they are neither eliminated  nor affected in the phagocytic capacity [47,48] by cyclophosphamide, while other cells such as lymphocytes and neutrophils are deleted .
Taken together, these results indicate that modulation of glucocorticoids action may be an important tool to eventually, develop strategies to reinstall or reinforce an active immune response in late sepsis patients to prevent, for example, opportunistic infections or viral reactivation .
In recent years, two different types of regulatory/suppressor cells have been also described as responsible for immunosuppression. Thus, myeloid-derived GR-1+CD11b+ suppressor cells and CD4+CD25+FoxP3+ regulatory T cells have been found to be involved in tumour models [39,40] in traumatic stress  as well as in polymicrobial sepsis models [28,35].
Consistent with these reports, in this study we observed an increase of these suppressor/regulatory cells in LPS-induced tolerant/immunosuppressed mice. Thus, taking into account the involvement of glucocorticoids in the tolerance/immunosuppression phenomenon, we analysed the possible relationship between glucocorticoids and these cell populations. Thus, considering that other authors have shown an induction of these cells by glucocorticoids in vitro[36,50], we extended these observations and demonstrated that dexamethasone was also capable of inducing an increase of GR-1+CD11b+ and CD4+CD25+FoxP3+ regulatory T cells in vivo.
In line with these observations, we designed experiments to elucidate the role of these cells in the establishment and maintenance of endotoxin tolerance.
The results showed that a significant decrease of GR-1+CD11b+ and CD4+CD25+FoxP3+ regulatory T cells by both cyclophosphamide and gemcitabine treatments did not modify the establishment or the maintenance of tolerance to LPS, suggesting that the presence of these cells is more an epiphenomenon induced by glucocorticoids than the cause of tolerance, although we do not discard that these cells could eventually be part of a redundant immunosuppressor mechanism. In line with this, Delano et al.  have demonstrated, in polymicrobial sepsis models, that expansion of GR-1+CD11b+ cells contributes to sepsis-induced T cell suppression, while Scumpia et al., although they have shown a relative increase in Treg numbers and increased suppressive activity, considered that these cells do not contribute to overall survival in this model .
Conversely, although the significantly diminished number of GR-1+CD11b+ and CD4+CD25+FoxP3+ cells observed in our report did not alter the tolerant state, direct evaluation of immunological status was not possible in this model, as both cyclophosphamide and gemcitabine, by themselves, can enhance or inhibit the immune response [51–53]. In fact, it is known that the effects of these drugs are not limited to these suppressor/regulatory cells as other cells, including lymphocytes T, B and neutrophils, could also be deleted or affected [49,54,55].
Endotoxin tolerance and Gram-negative sepsis seem to share pathways of immunosuppression mechanisms. Thus, the striking similarities observed between sepsis-induced macrophages dysfunction and endotoxin-tolerized macrophages , the common molecular mechanisms underlying hyporesponsiveness to LPS exhibited by monocytes from patients with sepsis and endotoxin tolerant cells [56,57] and the relationship between endotoxin tolerance and susceptibility to reinfection in patients with severe sepsis , are some of the similarities found in tolerance/sepsis-induced immunosuppression.
In brief, our results show that glucocorticoids play a crucial role in tolerance and/or immunosuppression induced by LPS, suggesting that the management of glucocorticoids could be a critical step for overcoming these effects. In addition, the presence of GR-1+CD11b+ and CD4+CD25+FoxP3+ regulatory T cells seems to be a peripheral phenomenon of glucocorticoid action in the establishment and maintenance of LPS tolerance. Conversely, despite the fact that LPS-induced tolerance is the initial step of immunosuppression in this model, direct evidence of these cells in immunosuppression is still lacking.
Whether our results can be extended to other states of immunosuppression induced by other agents is still not known, and will deserve further studies in the future.
We thank Dr Susana Fink for critical reading of the manuscript and Mr Antonio Morales for technical assistance. This work was supported by grants from Agencia Nacional de Promoción Científica y Tecnológica (PICT-2005-38197) and Fundación Alberto J. Roemmers.