Is Fas Ligand or Endotoxin Responsible for Mucosal Lymphocyte Apoptosis in Sepsis? FREE

DESPITE THE advent of a number of modern medical therapeutic approaches, sepsis and multiple organ failure remains one of the major causes of morbidity and mortality (as high as 60%) in the intensive care unit.¹ Many studies have suggested that development of multiple organ failure is strongly linked to the initial presentation of sepsis.² However, the precise mechanism by which sepsis induces subsequent multiple organ failure is still poorly understood. In this respect, over the last few years, an increasing number of studies have focused on the contribution of the barrier failure of the gut in the initiation and development of these conditions.³

Intestinal intraepithelial lymphocytes (IELs) are a subpopulation of the mucosa-associated lymphoid tissue. In the small intestine, IELs lie within the epithelial lining, which serves as the only layer to separate the intestinal lumen from the internal environment. These cells are, therefore, constantly exposed to foreign antigens and, thus, are thought to play a key role in the modulation of the mucosal immune system. The important role of IELs in gut immunity has been demonstrated in the maintenance of epithelial integrity,⁴ cytolytic potential,⁵ and the secretion of multiple cytokines in vitro.⁶ Studies have shown that the cytotoxicity of small-intestinal IELs is enhanced and CD4⁺ helper T cells are lost in mice and humans with inflammatory bowel disease.^{7 - 8} Such a loss of helper T-cell potential in response to localized inflammation might serve to compromise local immune function and directly and/or indirectly affect the function of the mucosal barrier of the gut. Interestingly, such changes in the gut's barrier function have been documented in sepsis and shock by a number of laboratories.^{9 - 10} However, the pathological mechanism for such changes and the role of local immune cells in this process are unclear. Over the past decade, IELs have received more attention by scientists because of their unique characteristics (ie, they contain 80%-90% T cells and most are CD8⁺ with an equal frequency of either γ δ or α β T-cell receptor) and divergent lineages (T cells develop with or without thymic influence).^{11 - 12} However, little is known about the mucosal, IEL immune response during sepsis or what mediators affect it.

Programmed cell death or apoptosis (A_o)is a cellular suicide program that is a regulated nonnecrotic form of cell death. For most cells, A_o usually is a constitutive response, but A_o can also be induced in certain immune cells such as T and B lymphocytes via immunologic and/or pathological processes.¹³ Apoptosis in activated T cells can be regulated by a variety of signals, of which the Fas (CD95):Fas ligand (FasL) pathway is best known.¹⁴ The induction of Fas/FasL appears to serve as an internal brake that restrains the antigen-activated T cell's response. However, A_o is also regulated by many stress mediators that are reported to be released during trauma and sepsis.² It has also been shown that intravenous administration of endotoxin to animals induces the release of a wide variety of mediators (eg, cytokines, hormones) that may account for circulatory and metabolic disturbances as well as immune dysfunction of the host. However, the contribution of these agents to the development of A_o and/or changes in mucosal immunity during sepsis are unclear.

Previous studies from our laboratory have indicated that a variety of immune cells from different compartments, such as thymocytes,¹⁵ bone marrow cells,¹⁵ Peyer's patch B cells,¹⁶ macrophages,¹⁷ and small-intestinal lamina propria mononuclear cells,¹⁸ undergo apoptosis following the onset of polymicrobial sepsis. The data from Peyer's patch and lamina propria lymphocytes suggest that increased A_o is associated with the activation of these cells during sepsis. However, there is little information concerning the response of IELs during sepsis on either a functional or a phenotypic level, as well as whether A_o is induced. The aim of this study, therefore, was to determine not only if polymicrobial sepsis causes changes in the functional (cytokine profile) and phenotypic state of the small-intestinal IELs, but if this was associated with the development of A_o. Furthermore, our aim was to determine not only what regulates phenotypic and/or apoptotic events, but also what impact this process has on the animal's survival.

Polymicrobial sepsis was induced in mice with the methods described by Baker et al¹⁹ and Ayala et al.²⁰ Male inbred C3H/HeNCrlBR (C3H/HeN endotoxin-sensitive) (Charles River Laboratories, Wilmington, Mass), C3H/HeJ (endotoxin-tolerant), and C3H/HeJ-FasL^gld (endotoxin-tolerant/Fas ligand–deficient) (Jackson Laboratory, Bar Harbor, Me) mice, 7 to 10 weeks old, were lightly anesthetized with methoxyflurane (Metofane, Pitman-Moore, Mundelein, Ill), shaved at the abdomen, and scrubbed with betadine. A midline incision (1.5-2.0 cm) was made below the diaphragm to expose the cecum. The cecum was ligated and punctured twice with a 22-gauge needle and gently compressed to extrude a small amount of cecal content through the punctured holes. The cecum was returned to the abdomen and the incision was closed in layers. Next, the mice were resuscitated with 0.8 mL of lactated Ringer solution by subcutaneous injection. For sham-controls, the cecum was neither ligated nor punctured. All procedures were carried out according to the National Institutes of Health Guidelines on Laboratory Animals and approved by the Animal Welfare Committee of Rhode Island Hospital, Providence.

The IELs were isolated using the method described by Davies and Parrott²¹ with minor modifications. Mice were killed 4 or 24 hours after CLP by methoxyflurane overdose and the small intestine was immediately removed and washed thoroughly with cold Hanks balanced salt solution (calcium free and magnesium free) (Gibco, Grand Island, NY). The Peyer's patches, fat, and mesentery were dissected and removed and the gut was slit open and cut into segments (0.5-1 cm). The gut segments were incubated for 90 minutes at 37°C in 3 changes of 30 mL of Hanks balanced salt solution containing 1-mmol/L dithiothreitol and 1-mmol/L EDTA (Sigma Chemical Co, St Louis, Mo) with continuous shaking in a water bath. The cell suspension from 3 incubations were pooled, centrifuged, resuspended, and filtered by passing through a loosely packed nylon wool column at room temperature to remove mucus, tissue debris, and dead cells. The cells were washed twice with RPMI complete medium with 2% fetal calf serum. Isolated IELs were further purified on a colloidal suspension (Percoll, Sigma) gradient.²¹ Eight milliliters of cells suspended in the 44% colloidal suspension was overlaid onto 5 mL of the 67% colloidal suspension in a 15-mL conical tube and centrifuged at 600g for 20 minutes at 4°C. Cells collected from the interface of the 44% and 67% colloidal suspensions were washed twice and resuspended at 10⁶ cells per milliliter of RPMI complete medium. The viability of the recovered cells was determined by trypan blue exclusion and was found to be greater than 90%.

To correlate the changes in the percentage of cells that were A_o positive with phenotypic expression, IELs were stained with a combination of phycoerythrine-conjugated anti-CD4 monoclonal antibody (mAb) (clone RM4-5, rat IgG2a), cychrome-conjugated anti-CD8α mAb (clone 53-6.7, rat IgG2a) (Pharmingen, San Diego, Calif), and fluorescein isothiocyanate (FITC)–labeled TUNEL (terminal deoxynucleotidyl transferase [TdT]–mediated deoxyuridine 5-triphosphate [dUTP] nick-end labeling) (Boehringer, Indianapolis, Ind), an in situ cell death detection kit, for detection of A_o.

Three-color flow cytometric analysis was carried out according to methods previously described by our laboratory.¹⁶ The analysis was performed on a flow cytometer (FACSort, Becton-Dickinson, San Jose, Calif). The FITC, phycoerythrine, and cychrome were excited at 488 nm with the argon laser. The FITC emission was detected at 525 nm; phycoerythrine at 575 nm; and cychrome at 670 nm. The overlapping emission of FITC, phycoerythrine, and cychrome was corrected by electronic compensation and no fewer than 10,000 cells were assessed in each sample. PC-Lysis Version 1.0 software (Becton-Dickinson) was used to establish 4 regions (phycoerythrine single-positive, cychrome single-positive, double-positive, and double-negative cells T cells) according to the fluorescent intensity of the nonspecific isotypic antibody controls using 2-color fluorescent cytograms. Histograms for cells from each region were created of the cell number vs the third color (eg, dUTP-FITC stain) fluorescent intensity. From the histogram, the percentage of apoptotic cells was determined based on the positive (DNase-treated) and negative (DNase-untreated) sham-controls.

Total cellular RNA from IELs was isolated using an RNA isolation reagent (Tripure, Boehringer-Mannheim GmBH, Mannheim, Germany) following the manufacturer's instructions, suspended in RNase-free water, quantitated spectrophotometrically at 260 nm, and stored at −70°C.

An anti-sense complementary RNA multiprobe used for RPA was synthesized from a complementary DNA template set (mCK-1, Pharmingen) with [³²P]-αUTP (NEN, DuPont, Boston) using an in vitro transcription kit according to the standard protocol from the manufacturer (Pharmingen).¹⁸ This mCK-1 template set contains fragments that correspond to mouse cytokines genes IL-2, IL-4, IL-5, IL-6, IL-9, IL-10, IL-13, and IL-15 and IFN-γ, as well as the housekeeping genes L32 and GAPDH.

The RPA was performed according to the manufacturer's instructions (Pharmingen). The RNA samples were hybridized with the anti-sense RNA probe at 56°C for 12 to 16 hours. Following hybridization, free probe and unprotected RNA were digested with an RNase solution (RNase A + RNase T1). The double-stranded RNA was then extracted (chloroform–isoamyl alcohol), precipitated (ethanol), and separated in a denaturing 8-mol/L urea–6% polyacrylamide gel. A lane with the undigested probe set was to serve as size markers. The gel was dried, exposed to autoradiography film (NEN), and the films developed in processors (M35 and M35A X-OMAT, Kodak, Rochester, NY).

A standard curve plotted with the undigested probe markers was used to identify the bands of various cytokine genes in the experimental samples. Quantitation of messenger RNA (mRNA) was performed by densitometric scanning of the autoradiogram (MOCHA Image Analysis/SigmaGel work station, Jandel Sci, Corte Madera, Calif). To quantify the individual cytokine mRNA levels, 2 housekeeping gene (L32 and GAPDH) transcripts were also included and the sample values were corrected (percentage of housekeeping gene) to the extent of expression of the housekeeping gene.

The data are presented as mean ± SEM for each group. Differences in percentile data (ie, percentage A_o positive) were considered to be significant at P<.05, as determined using the Mann-Whitney U test. The survival data were compared using the Fisher exact test and considered to be significant at P<.05.

Total viable cell yield of IELs from the small intestine of C3H/HeN mice was markedly decreased in mice with sepsis at 4 hours (sham-controls, 1.3 ± 0.1 × 10⁷ cells vs CLP, 0.9 ± 0.2 × 10⁷ [P<.05; n = 6 mice per group]) but not at 24 hours (sham-controls, 1.2 ± 0.3 × 10⁷ vs CLP, 1.3 ± 0.2 × 10⁷ cells) following the onset of sepsis. To further characterize the phenotypes of isolated IELs from C3H/HeN mice, cells were stained with mAbs against the T-cell surface markers CD4 and CD8. Figure 1 illustrates the percentage of cells positively stained with anti-CD4 and anti-CD8 in IELs from sham-controls and CLP mice. A slight but significant decline of CD8 single-positive, CD4⁻CD8⁺, cells was observed at both 4 (Figure 1, top) and 24 (Figure 1, bottom) hours after CLP. However, the percentage of CD4 single-positive, CD4⁺CD8⁻ and CD4⁺CD8⁺, cells was not markedly decreased until 24 hours after CLP, while the percentage of CD4⁻CD8⁻ cells was significantly increased at 24 hours.

To correlate the frequency of A_o with phenotypic expression, IELs isolated from C3H/HeN mice were stained with a combination of anti-CD4 and anti-CD8 mAbs for phenotypic characterization and TUNEL assay for assessment of A_o. Figure 2 shows the percentage of apoptotic cells stained with CD4 and/or CD8 markers in sham-controls and CLP mice at 4 (Figure 2, top) and 24 hours (Figure 2, bottom) following the onset of sepsis. At 24 but not at 4 hours, a significant increase in the frequency of A_o in cells expressing CD4⁺CD8⁻, CD4⁻CD8⁺, and CD4⁻CD8⁻ was observed in the septic mice. However, there were no important differences in the CD4⁺CD8⁺ population between sham-controls and septic mice.

Expression of cytokine mRNA was analyzed by RPA. The results of cytokine gene expression from freshly isolated (unstimulated) IELs are shown in Figure 3. Using the commercial multiprobe set, IL-2, IL-10, IL-15, and IFN-γ genes were consistently detected in both sham-controls and CLP mice, while signals for IL-4, IL-5, IL-6, IL-9, and IL-13 genes (also included in the multiprobe set) were not detected in our system. Quantitative analysis of the expressed cytokine mRNA levels was estimated in the autoradiographic intensity normalizing to housekeeping gene (L32 or GAPDH) expression. Among the detected cytokine genes in IELs, there was a marked increase in IL-2, IL-15, and IL-10 mRNA evident 24 (Figure 3, bottom) but not 4 hours (Figure 3, top) after CLP. However, expression of IFN-γ mRNA gene did not differ between sham-controls and CLP mice at either time point.

Because recent studies have indicated that Fas antigen and FasL may mediate A_o seen to be associated with the activation of lymphocytes during an immune response,²² it was our hypothesis that the increase in lymphocyte A_o seen here to be associated with increased cytokine expression might be reflective of the Fas/FasL-mediated process. To examine this, we conducted experiments to assess the frequency of A_o with phenotypic expression in IELs at 24 hours after the onset of sepsis using a mutant mouse strain with a defective FasL gene gld (C3H/HeJ-FasL^gld)²³ and C3H/HeJ mice (as the endotoxin-tolerant background control). The percentages of both CD4⁺CD8⁻ and CD4⁻CD8⁺ cells were markedly decreased, while a marked increase in CD4⁻CD8⁻ cells was seen in septic C3H/HeJ mice at 24 hours (Figure 4, top). However, these changes were not observed in C3H/HeJ-FasL^gld mice (Figure 4, bottom). A significant increase in the frequency of A_o in cells expressing CD4⁺CD8⁻ and CD4⁻CD8⁺ was observed in C3H/HeJ mice 24 hours after CLP (Figure 5, top). Interestingly, all of the IELs' phenotypic populations showed a slight trend toward lower apoptotic frequency in septic C3H/HeJ-FasL^gld mice although these were not statistically significant (Figure 5, bottom).

To the extent the FasL deficiency might provide any protection against septic mortality, C3H/HeJ and C3H/HeJ-FasL^gld mice were subjected to the CLP procedure and monitored for their survival over a 7-day period. Figure 6 shows that although the survival rates of 2 mouse strains were not significantly different from each other until day 7, the latent and lower mortality rates were observed in C3H/HeJ-FasL^gld mice consistently on days 2 through 7.

The murine small-intestinal IELs are a heterogeneous mononuclear cell population that contains approximately 80% to 90% CD3⁺ T cells.²⁴ Most IEL T cells are CD4⁻CD8⁺ (approximately 50%-75%) and contain large intracytoplasmic granules of perforin and serine esterases, the concentration of which are increased on antigenic activation.^{12 ,25} These large granular lymphocytes are unusual for a T-cell population and are thought to mediate natural killer cell cytotoxic activity.²⁶ Total viable IEL yield in our preparation from C3H/HeN mice was markedly decreased at 4 but not 24 hours after CLP. However, it should be noted that although at 24 hours total IEL yield did not differ between sham-control and septic animals, a marked increase in CD4⁻CD8⁻ cells was observed in septic mice, which may be reflective of a route of entry for neutrophil, macrophage, and/or other non–T-cell populations. Thus, it appears that localized immune cell activation and/or the recruitment of inflammatory cells to the gut occurs during sepsis.

To assess the effect of polymicrobial sepsis on the T-cell subpopulations of IELs, the phenotypic expression was examined using 3-color flow cytometric analysis. Besides CD4⁻CD8⁺ T cells, IELs typically consist of about 7.5% CD4⁺CD8⁻ and approximately 10% double-positive and 7.5% double-negative T cells.²⁷ The percentage of CD4⁺CD8⁻ and CD4⁻CD8⁺ T-cell subpopulations in the IEL preparations from sham-control C3H/HeN mice was consistent with those reported in the literature.²⁷ Interestingly, only the percentage of CD4⁻CD8⁺ cells was markedly decreased in the septic mice at both 4 and 24 hours after CLP, while the other cell populations did not differ until 24 hours following the onset of sepsis. A similar finding was made in C3H/HeJ but not in C3H/HeJ-FasL^gld mice. Thus, we conclude that sepsis produces an early effect on the cytotoxic T-cell population that is not evident until late in the helper T cell phenotype.

The IELs, because of their anatomical location, are thought to play a critical role in the mucosal immune system through the surveillance of the local antigen environment. Although little is known about the potential function of IELs in vivo, many in vitro studies have been conducted to elucidate the biological function in this population. Consisting primarily of T cells, of which, approximately 70% are CD8⁺ cytotoxic T cells, IELs have been shown to possess both a competent tumor and microbial cytotoxic capacity.⁵ The IELs have also been shown to contribute to the maintenance of epithelial cell integrity,⁴ and they also appear to be involved in the maintenance of mucosal secretory IgA immune response.²⁸ Furthermore, IELs have been shown to secrete a variety of cytokines on stimulation and/or activation.⁶ In this regard, our study demonstrates that freshly isolated small-intestinal IELs from septic mice express higher levels of mRNA for IL-2, IL-10, and IL-15 genes but not for the IFN-γ gene. While these changes seem to be relatively small, we speculate that this is in part a result of the extended period of time required to isolate these ex vivo cells. These results, based on cytokine gene expression, also suggest that there seems to be no clear in vivo preference toward Th1 or Th2 phenotype in murine IELs. These findings are in agreement with those of Clevers et al,²⁹ who showed that IELs from mice contain a mixture of Th1- and Th2-type cells. However, the increased release of these cytokines, ie, interleukin (IL) 2, IL-15, suggests that these cells have most likely been activated in response to stimuli associated with the onset of sepsis. In this regard, the induction of T-cell activation sets in motion not only processes involved with the development of a cell-mediated immune response (such as lymphokine release, major histocompatibility complex antigen expression), but it also induces mechanisms to eventually contain such a response, via up-regulation of expression of Fas/FasL¹⁴ as well as other receptors-pathways that can induce apoptotic cell death (via autocrine, paracrine, and/or exocrine cytokine-mediator pathways).³⁰

Apoptosis is a constitutive process for most of the cells that carry out their own death through activation of an internally encoded suicide program and initiate A_o.³¹ However, A_o can also be induced in some immune cells through the antigenic engagement of the T-cell or B-cell receptor complex (ie, activation-induced cell death)³² or by various factors (eg, glucocorticoids, inflammatory cytokines, nitric oxide) released during pathological insult. Studies from our laboratory^{15 - 17} and others^{33 - 34} indicated that increased A_o is evident in a variety of immune cell types (T and B cells, macrophages, and granulocytes) from different organs and tissues including the thymus, bone marrow, spleen, Peyer's patches, and lamina propria mononuclear cells following the onset of sepsis. Our results show that the apoptotic frequency of T lymphocytes in intestinal IEL preparation from endotoxin-sensitive C3H/HeN mice was increased in septic mice. A marked increase in the frequency of A_o in CD4⁺CD8⁻, CD4⁻CD8⁺, and CD4⁺CD8⁺ T cells was evident 24 hours after CLP. This indicates that induction of A_o is an event that is not evident in this cell population until the late hypodynamic-hypometabolic stage of polymicrobial sepsis, a period in which organ perfusion and oxygenation are expected to be compromised in the gut.³⁵ These findings are in keeping with our previous findings in B cells from the Peyer's patch and bone marrow, in which a marked increase in apoptotic frequency was not detected until the late time point (24 hours after CLP).^{15 - 16} However, the results differ from the thymus,¹⁵ peritoneal macrophage,¹⁷ and lamina propria mononuclear cells¹⁸ in which the onset of A_o was detected as early as 4 hours after CLP.

It has been suggested that activated T cells up-regulate the expression of the Fas/FasL gene, and undergo activation-induced A_o. Fas ligand (FasL or CD95L, members of the tumor necrosis factor superfamily) interacts with their receptors, Fas (APO-1 or CD95, members of the tumor necrosis factor receptor superfamily), to regulate important functions in immune cell activation, differentiation, proliferation, and survival.¹⁴ Endotoxin (a key mediator of gram-negative bacteremia) has been shown to induce thymic A_o³⁶ ; however, the dosage of lipopolysaccharide required to induce marked A_o was 50 µg or higher per mouse (intraperitoneally; equivalent to 2.5 mg of lipopolysaccharide per kilogram of body weight). This concentration of lipopolysaccharide typically produces blood levels of endotoxin that are considerably higher than the circulating levels of this agent detected in patients³⁷ or mice²⁰ with sepsis. Furthermore, previous studies from our laboratory have demonstrated that genetically related strains of mice, C3H/HeN (endotoxin-sensitive strain) and C3H/HeJ (endotoxin-tolerant strain) showed no difference in survival rate when they were subjected to CLP.³⁸ In this respect, we attempted to determine whether Fas/FasL and/or endotoxin contribute to the increased A_o in IELs during sepsis. C3H/HeJ (endotoxin-tolerant) and C3H/HeJ-FasL^gld (a mouse mutant strain of endotoxin-tolerant/FasL-deficient) mice were subjected to CLP and the frequency of A_o in IELs was determined. A marked increased A_o in CD4⁺CD8⁻ and CD4⁻CD8⁺ cells was observed in endotoxin-tolerant C3H/HeJ mice. Alternatively, IEL T cells from C3H/HeJ-FasL^gld showed a slight (though not significant) decrease in A_o after CLP. These findings, therefore, indicate that FasL but not endotoxin mediates the induction of A_o in IELs, since sepsis causes IEL T-cell A_o in both endotoxin-sensitive and -tolerant mice. This is in keeping with the studies reported by Hiramatsu et al,³⁴ who showed by in situ TUNEL that CLP induced A_o in different organs including the gut by an endotoxin-independent pathway.

In light of this, we speculated that Fas/FasL–mediated A_o may play an important role in morbidity and/or mortality during sepsis. We, therefore, attempted to determine the relationship between FasL deficiency to the mouse's survival by subjecting C3H/HeJ and C3H/HeJ-FasL^gld mice to CLP and comparing their mortality. A significantly lower mortality rate was observed in FasL-deficient mice after CLP. On the surface, this would seem to be in keeping with the hypothesis that an overzealous gut inflammatory response associated with increased Fas/FasL-mediated A_o may be contributing to not only the dysfunction of the septic gut barrier but also to the associated morbidity/mortality seen in sepsis. However, although it is tempting to speculate that the protective effects on survival seen here might be related to the decreased A_o in the gut IELs, we recognize that these protective effects of FasL deficiency may be directed at other immune tissues or cells (eg, phagocytes that can act as a source of FasL) and/or at nonimmune cells or tissues that also respond to the FasL gene product.³⁹

John C. Marshall, MD, Toronto, Ontario: This work is intriguing and important in 2 separate domains. One is the inference that the gut immune system potentially has a role in survival in a model of intra-abdominal infection. The second is that you have shown us an endotoxin-independent and Fas-dependent pathway to mortality in a classic model of intra-abdominal sepsis.There are wonderful possibilities for that sort of work in the future. I have 3 questions that I wanted to ask you.

The first one is that in your manuscript you raise the possibility that when you see an increased number of double-negative cells in the gut wall that you may be looking at infiltration by nonlymphocytes, and I wonder if you have done any studies to see if there are increased numbers of neutrophils in the gut; in other words, to see whether or not the intraepithelial lymphocyte population is suppressing inflammation and you are losing a suppressive influence?

Second, I am wondering if you have looked at all at what might be responsible for the changes in phenotype and susceptibility to apoptosis? Is it a flow-dependent phenomenon? If you don't resuscitate the animals, do you see a worsening of it? Or is it secondary to changes in the bacterial flora, since we know that the bacterial flora of the small intestine will change during cecal ligation and puncture.

And then finally, I wanted to ask you with regard to the survival studies you have done, whether you have done histology. This to my knowledge is the first model that I have seen that really reflects what happens in human multiple organ failure, which is that mortality is not early but it is late, and the fact that your survival curves diverge at about 7 days is most intriguing. I am wondering if you have looked to see if there is increased evidence of inflammation in the liver or the lungs of the HeJ but not GLD mice?

Dr Chung: Thank you, Dr Marshall. With regard to your first question about double-negative cell population increase, it's possible that neutrophil and other cells are from other compartments. We haven't looked at the neutrophil phenotype due to some limitation of the cell yield. The cell yield sometimes is pretty low.

And the second question about does the phenotype reflect the apoptosis, is there a relationship with some resuscitation or the flow? We haven't seen that; we haven't looked at that yet. That is a good suggestion.

And the third question is the survival. That is another excellent suggestion. We should see the inflammation of the HeJ mice. Thank you.

John A. Mannick, MD, Boston, Mass: If I understand previous work from your laboratory correctly, you have looked in the gut before for apoptosis in your model of trauma and hemorrhage and found there that the cells dying were in fact B cells rather than T cells. Am I correct? Did you look at what is happening to B cells in the system you are now looking at, namely, the septic system, because of course B cells might be the truly proximate source of the IgA in the gut. Are they dying as well?

Dr Chung: Yes. We have another set of experiments dealing with the lamina propria lymphocytes, and then we did see the B cells increase the apoptosis after the CLP at both time points, at 4 and 24 hours.

This investigation was supported by grant R01-GM53209 from the National Institutes of Health Bethesda, Md.

Presented at the 18th Annual Meeting of the Surgical Infection Society, New York, NY, April 30, 1998.

We would like to thank Sally A. Johnson and Paul Monfils at the Core Laboratories Facilities at Rhode Island Hospital, Providence, for their assistance with the flow cytometry.

Corresponding author: Alfred Ayala, PhD, Center for Surgical Research, 211 Middle House, Rhode Island Hospital, 593 Eddy St, Providence, RI 02903 (e-mail: AAYALA@RIHOSP.EDU).

Over the last few years, it has become apparent that the intestine as an immunological organ may play not only an important role in the maintenance of gut barrier function but in the evolution of multiple organ failure associated with shock and sepsis. However, little is known about the pathobiologic characteristics of these changes and what regulates them. We, therefore, have attempted to examine the effect of sepsis on the phenotypic as well as functional response of one compartment of the mucosal immune system, the IEL. Our findings indicate that sepsis induces an increase in A_o in different populations of IELs that include CD4⁺CD8⁻, CD4⁻CD8⁺, CD4⁺CD8⁺, and/or CD4⁻CD8⁻ T lymphocytes. A marked increase in IL-2, IL-10, and IL-15 mRNA gene expression in IELs was observed 24 hours after the onset of sepsis. These results suggest that the increase in A_o in intestinal IELs is associated with changes in cytokine gene expression following the onset of sepsis; there is localized immune cell activation leading to activation-induced apoptotic cell death during sepsis; and this IEL apoptotic response mediated by FasL and FasL deficiency seems to provide a survival advantage during sepsis. Thus, the use of agents that inhibit FasL during sepsis might not only be useful in preventing the potential pathological effects of sepsis-induced gut lymphoid A_o but the associated septic mortality.