[FABP5 promotes cell growth, invasion and metastasis in cervical cancer]

Objective: To investigate the functions of FABP5 in the carcinogenesis and development of cervical cancer. Methods: The expression of FABP5 was detected in several cervical cancer cell lines (C33A, Siha, Caski, HeLa and HCC94), 206 cases of cervical cancer tissues with stage Ⅰa2-Ⅱa2 and 40 cases of normal cervical tissues by real-time PCR and Western blotting. Then, the cells were infected with lentivirus-mediated siRNA-targeting FABP5. CCK-8 cell proliferation, colony formation, wound healing and transwell assays were used to investigate the effects of FABP5 on in vitro cell proliferation, migration and invasion. And in vivo xenograft model and lung metastasis model were used to observe the transplanted tumor growth and metastasis in female athymic nude mice. Furthermore, the total protein and RNA were extracted from the primary xenografts to determine the expression levels of FABP5, metalloproteinase-2 and metalloproteinase-9 using Enzyme linked immunosorbent assay (ELISA), real-time PCR and Western blotting. Results: FABP5 expression was found to be significantly unregulated in cervical cancer tissues than that in normal cervical tissues (P<0.05). Compared with the Siha-NC group and uninfected group, the expression of FABP5 mRNA and protein in Siha-FABP5-RNAi group was significantly inhibited along with the decrease of cell proliferation, colony formation, wound healing and invasion ability. The clone formation rates of Siha cells in uninfected group, Siha-NC group and Siha-FABP5-RNAi group were (84.6±4.5)%, (84.6±5.1)% and (21.2±2.6)%, respectively. Moreover, the transwell assay showed that invasive cells in three groups were (72.8±4.7)/HPF, (72.6±3.3)/HPF and (21.4±2.3)/HPF, respectively. All of the difference was statistically significant (P<0.05). Furthermore, FABP5 silencing significantly reduced tumor growth and lung metastases in nude mice in vivo (P<0.001). The subcutaneously xenografted volume in uninfected group, Siha-NC group and Siha-FABP5-RNAi group was (921.4±63.0) mm(3,) (1 021.4±56.0) mm(3) and (139.6±36.0) mm(3,) respectively. The real-time quantitative PCR results showed that the relative expression levels of MMP-2 and MMP-9 mRNA were 1.00±0.10 and 1.00±0.10, 1.00±0.10 and 1.00±0.10 as well as 0.34±0.13 and 0.38±0.17 in xenografted tumor tissues of uninfected group, Siha-NC group and Siha-FABP5-RNAi group, respectively. MMP-2 and MMP-9 was significantly downregulated after FABP5 inhibition(P<0.05). Additionally, the protein expression trend of MMP-2 and MMP-9 in three groups was consistent with the mRNA levels. Conclusion: FABP5 might promote the carcinogenesis and metastasis of cervical cancer via up-regulating MMP-2 and MMP-9.

Role of selenoprotein S (SEPS1) -105G>A polymorphisms and PI3K/Akt signaling pathway in Kashin-Beck disease

OBJECTIVE

To investigate the relationship between SEPS1 polymorphism and phosphatidylinositol 3-kinase (PI3K)/Akt signaling pathway in Kashin-Beck disease (KBD) and further explore the pathogenesis of KBD.

METHODS

Polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP) was used to detect SEPS1 -105G>A polymorphism in 232 cases and 331 controls. The protein expressions of PI3K/Akt signaling molecules in whole blood and chondrocytes were detected by Western blot.

RESULTS

The frequencies of SEPS1 -105G>A genotype AA (21.1% vs 3.0%) and minor allele A (34.1% vs 16.0%) in KBD are significantly higher than those in controls (OR: 8.020, 95% confidence interval (95% CI) 6.341-10.290, P < 0.0001; OR: 2.470, 95% CI 2.001-4.463, P < 0.0001, respectively). SEPS1 AA genotype was an independent risk factor for KBD (adjusted OR: 9.345, 95% CI 4.254-20.529; P < 0.0001). The expression of Gβγ, PI3Kp110, pAkt and pGSK3β in KBD group were higher than that in control group (all P < 0.05). Gβγ, pAkt and pGSK3β protein expression of AA and GA increased than GG (all P < 0.05). Cell apoptosis was increasing and molecule expression of PI3K/Akt signaling pathway were up-regulated in the tert-Butyl hydroperoxide (tBHP)-injured group, the cell apoptosis and expression levels of PI3K/Akt in Na2SeO3 group were decreased.

CONCLUSIONS

The SEPS1 -105G>A is associated with an increased risk of KBD and influences the expression of PI3K/Akt signaling pathway in KBD patients. Apoptosis induced by tBHP in chondrocyte might be mediated via up-regulation of PI3K/Akt, Na2SeO3 has an effect of anti-apoptosis by down-regulating of PI3K/Akt signaling pathway.

Molecular characterization, expression pattern and association analysis of the porcine BTG2 gene

B-cell translocation gene 2 (BTG2), a member of the B-cell translocation gene family with anti-proliferative properties, have been characterized to be involved in cell growth, differentiation and survival. In this study, we cloned the full length sequences of cDNA and genomic DNA of BTG2 gene from the porcine skeletal muscle. Spatial expression analysis showed that the porcine BTG2 gene is expressed predominantly in muscle. Temporal expression analysis in longissimus dorsi muscle demonstrated that the expression of BTG2 gene has the highest expression at 60 days old in Large White while with a peak expression at 120 days old in Meishan. Temporal analysis also revealed that the expression of BTG2 gene is generally higher in Large White than in Meishan at all the developmental stages tested (65 days of conception and 3, 35, 60, 120, and 180 days of postnatal). A single nucleotide polymorphism (G417C) in the intron of BTG2 gene was then detected by PCR-RFLP in Large White × Meishan F2 resource population and association analysis suggested that this polymorphic site had significant association (P < 0.05) with the buttock fat thickness, fat percentage, lean muscle percentage, ratio of lean to fat and carcass length.

Association study between polymorphisms in selenoprotein genes and susceptibility to Kashin-Beck disease

OBJECTIVES

Kashin-Beck disease (KBD) is a disabling osteoarthropathy involving growth cartilage endemic to selenium (Se)-deficient regions in China. Associations between genetic variation in selenoprotein genes and susceptibility to many diseases have recently been investigated but few studies have been performed on KBD. We found four genetic polymorphisms in selenoprotein genes and assessed their association with increased susceptibility to KBD.

METHODS

Four polymorphisms including GPX1 (rs1050450), TrxR2 (rs5748469), SEPP1 (rs7579) and DIO2 (rs225014) were analyzed for 161 KBD patients and 312 controls using polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP) or tetra-primer amplification refractory mutation system PCR (Tetra-primer ARMS PCR). Glutathione peroxidase (GPX) activity in whole blood was measured using a GPX assay kit. The mRNA expression of GPX1, nuclear factor-kappaB (NF-kappaB) p65 and p53 in both whole blood and articular cartilage tissue were detected using Real-Time PCR.

RESULTS

The genotypic and allelic frequency of GPX1 Pro198Leu was significantly different between KBD patients and controls (P=0.013, P=0.037). A significant increased KBD risk was observed in individuals with Pro/Leu or Leu/Leu (odds ratio=1.781; 95% confidence interval: 1.127-2.814) compared with Pro/Pro. No association was observed between the other three single nucleotide polymorphisms (SNPs) and KBD risk. In addition, GPX enzyme activity in whole blood was lower in the KBD group (P<0.01), and the GPX activity in whole blood decreased significantly in a subgroup of individuals representing Pro/Leu and Leu/Leu compared to Pro/Pro (P<0.01). In whole blood and articular cartilage tissue samples of KBD patients, GPX1 and NF-kappaB p65 mRNA levels were lower (P<0.01) while p53 levels were higher (P<0.001).

CONCLUSION

GPX1 Pro198Leu is a potential genetic risk factor in the development of KBD and the GPX1 Leu allele is significantly associated with higher KBD risk among the Chinese Han population and with lower GPX enzyme activity. The expression of apoptosis related molecules in KBD patients significantly differs from controls.

Proteolysis and class I major histocompatibility complex antigen presentation

The class I major histocompatibility complex (MHC class I) presents 8-10 residue peptides to cytotoxic T lymphocytes. Most of these antigenic peptides are generated during protein degradation in the cytoplasm and are then transported into the endoplasmic reticulum by the transporter associated with antigen processing (TAP). Several lines of evidence have indicated that the proteasome is the major proteolytic activity responsible for generation of antigenic peptides--probably most conclusive has been the finding that specific inhibitors of the proteasome block antigen presentation. However, other proteases (e.g. the signal peptidase) may also generate some epitopes, particularly those on certain MHC class I alleles. The proteasome is responsible for generating the precise C termini of many presented peptides, and appears to be the only activity in cells that can make this cleavage. In contrast, aminopeptidases in the cytoplasm and endoplasmic reticulum can trim the N terminus of extended peptides to their proper size. Interestingly, the cellular content of proteases involved in the production and destruction of antigenic peptides is modified by interferon-gamma (IFN-gamma) treatment of cells. IFN-gamma induces the expression of three new proteasome beta subunits that are preferentially incorporated into new proteasomes and alter their pattern of peptidase activities. These changes are likely to enhance the yield of peptides with C termini appropriate for MHC binding and have been shown to enhance the presentation of at least some antigens. IFN-gamma also upregulates leucine aminopeptidase, which should promote the removal of N-terminal flanking residues of antigenic peptides. Also, this cytokine downregulates the expression of a metallo-proteinase, thimet oligopeptidase, that actively destroys many antigenic peptides. Thus, IFN-gamma appears to increase the supply of peptides by stimulating their generation and decreasing their destruction. The specificity and content of these various proteases should determine the amount of peptides available for antigen presentation. Also, the efficiency with which a peptide is presented is determined by the protein's half life (e.g. its ubiquitination rate) and the sequences flanking antigenic peptides, which influence the rates of proteolytic cleavage and destruction.

Distinct Proteolytic Processes Generate the C and N Termini of MHC Class I-Binding Peptides

Most of the MHC class I peptides presented to the immune system are generated during the course of protein breakdown by the proteasome. However, the precise role of the proteasome, e.g., whether this particle or some other protease generates the carboxyl (C) and amino (N) termini of the presented 8- to 10-residue peptides, is not clear. Here, we show that presentation on Db of ASNENMETM, a peptide from influenza nucleoprotein, and on Kb of FAPGNYPAL, a peptide from Sendai virus nucleoprotein, was blocked by the proteasome inhibitor, lactacystin. Using plasmid minigene constructs encoding oligopeptides of various lengths, we found that presentation of ASNENMETM from C-terminally extended peptides that contain this antigenic peptide plus three or five additional amino acids and presentation of FAPGNYPAL from a peptide containing FAPGNYPAL plus one additional C-terminal residue required the proteasome. In contrast, the proteasome inhibitor did not reduce presentation of cytosolically expressed ASNENMETM or FAPGNYPAL or N-terminally extended versions of these peptides, suggesting involvement of aminopeptidase(s) in trimming these N-extended variants. Accordingly, when the N termini of these 3N-extended peptides were blocked by acetylation, they were resistant to hydrolysis by cellular aminopeptidases and pure leucine aminopeptidase. Moreover, if introduced into the cytosol, Ag presentation of these peptides occurred to a much lesser extent than from their nonacetylated counterparts. Thus, the proteasome is essential for the generation of ASNENMETM and FAPGNYPAL peptides from the full-length nucleoproteins. Although it generates the C termini of these presented peptides, distinct aminopeptidase(s) can trim the N termini of these presented peptides to their proper size.

Distinct proteolytic processes generate the C and N termini of MHC class I-binding peptides

Most of the MHC class I peptides presented to the immune system are generated during the course of protein breakdown by the proteasome. However, the precise role of the proteasome, e.g., whether this particle or some other protease generates the carboxyl (C) and amino (N) termini of the presented 8- to 10-residue peptides, is not clear. Here, we show that presentation on Db of ASNENMETM, a peptide from influenza nucleoprotein, and on Kb of FAPGNYPAL, a peptide from Sendai virus nucleoprotein, was blocked by the proteasome inhibitor, lactacystin. Using plasmid minigene constructs encoding oligopeptides of various lengths, we found that presentation of ASNENMETM from C-terminally extended peptides that contain this antigenic peptide plus three or five additional amino acids and presentation of FAPGNYPAL from a peptide containing FAPGNYPAL plus one additional C-terminal residue required the proteasome. In contrast, the proteasome inhibitor did not reduce presentation of cytosolically expressed ASNENMETM or FAPGNYPAL or N-terminally extended versions of these peptides, suggesting involvement of aminopeptidase(s) in trimming these N-extended variants. Accordingly, when the N termini of these 3N-extended peptides were blocked by acetylation, they were resistant to hydrolysis by cellular aminopeptidases and pure leucine aminopeptidase. Moreover, if introduced into the cytosol, Ag presentation of these peptides occurred to a much lesser extent than from their nonacetylated counterparts. Thus, the proteasome is essential for the generation of ASNENMETM and FAPGNYPAL peptides from the full-length nucleoproteins. Although it generates the C termini of these presented peptides, distinct aminopeptidase(s) can trim the N termini of these presented peptides to their proper size.

Matching antibody class with pathogen type and portal of entry: cognate mechanisms regulate local isotype expression patterns in lymph nodes draining the respiratory tract of mice inoculated with respiratory viruses, according to virus replication competence and site of inoculation

Intranasal deposition of Sendai virus (SV) in C57BL/6 mice provokes an Ab-forming cell (AFC) reaction in mediastinal (MLN) and cervical lymph nodes (CLN), which drain the lungs and upper respiratory tract, respectively. While the majority of AFC elicited by infectious SV at both sites produced IgG, the CLN response to SV rendered inactive in replication was restricted almost entirely to IgA, although isotype switching in mediastinal continued to be skewed heavily to IgG. However, in vitro restimulation of the accompanying virus-specific T cell populations from the two sites did not reveal any significant difference in lymphokine output, and isotype expression was not altered substantially in mice lacking IL-4 or IL-6 genes. To dissociate the response to specific Ags from the inflammatory reaction to viral infection, we examined the response to inactivated SV in the face of infection with influenza virus A/HKx31. The magnitude and IgA dominance of the anti-SV AFC population in the CLN were unaffected by a simultaneous, vigorous, IgG-dominated CLN anti-influenza reaction. Evidently, the characteristics of this antiviral response are determined primarily by cognate interactions. Moreover, the IgA bias of the CLN AFC response to inactivated SV was observed only when the virus was delivered intranasally: injection under the epidermis of the cheek, a site that has a lymphatic drainage into the CLN, resulted in an IgG-dominated CLN AFC reaction, lacking IgA. The site of deposition of a vaccine can thus have more influence on the pattern of isotypes induced than the site at which the immune response is initiated.

Matching antibody class with pathogen type and portal of entry: cognate mechanisms regulate local isotype expression patterns in lymph nodes draining the respiratory tract of mice inoculated with respiratory viruses, according to virus replication competence and site of inoculation

Intranasal deposition of Sendai virus (SV) in C57BL/6 mice provokes an Ab-forming cell (AFC) reaction in mediastinal (MLN) and cervical lymph nodes (CLN), which drain the lungs and upper respiratory tract, respectively. While the majority of AFC elicited by infectious SV at both sites produced IgG, the CLN response to SV rendered inactive in replication was restricted almost entirely to IgA, although isotype switching in mediastinal continued to be skewed heavily to IgG. However, in vitro restimulation of the accompanying virus-specific T cell populations from the two sites did not reveal any significant difference in lymphokine output, and isotype expression was not altered substantially in mice lacking IL-4 or IL-6 genes. To dissociate the response to specific Ags from the inflammatory reaction to viral infection, we examined the response to inactivated SV in the face of infection with influenza virus A/HKx31. The magnitude and IgA dominance of the anti-SV AFC population in the CLN were unaffected by a simultaneous, vigorous, IgG-dominated CLN anti-influenza reaction. Evidently, the characteristics of this antiviral response are determined primarily by cognate interactions. Moreover, the IgA bias of the CLN AFC response to inactivated SV was observed only when the virus was delivered intranasally: injection under the epidermis of the cheek, a site that has a lymphatic drainage into the CLN, resulted in an IgG-dominated CLN AFC reaction, lacking IgA. The site of deposition of a vaccine can thus have more influence on the pattern of isotypes induced than the site at which the immune response is initiated.

Gamma interferon is not essential for recovery from acute infection with murine gammaherpesvirus 68

Murine gammaherpesvirus 68 (MHV-68) when administered intranasally induces high levels of gamma interferon (IFN-gamma) in the lymphoid tissues of infected mice. In order to investigate the role of this cytokine in the immune response to MHV-68, mice which were congenitally deficient in the IFN-gamma gene (IFN-gamma knockout mice) were infected with the virus. Comparison of the courses of the disease in wild-type control and IFN-gamma knockout mice revealed surprisingly little difference. Both groups of mice had cleared infectious virus from the lungs 15 days after infection, although there did appear to be a slight delay in viral clearance in the IFN-gamma knockout mice. In addition, after the initial phase of viral clearance, the lungs of both groups remained clear of replicating virus throughout the course of the experiment, which concluded 34 days after infection. Consistent with these observations, cytotoxic T-cell activities were similar in the two groups of mice. Levels of latent virus were comparable in wild-type and knockout mice over the time course studied. Furthermore, analysis of the numbers, types, and activation status of cells in the lungs, lymph nodes, and spleens of control and knockout mice revealed no striking difference. This suggests that IFN-gamma is not essential for regulating the cell recruitment or proliferation that normally occurs during this viral infection. Apart from the expected lack of IFN-gamma, cytokine profiles were not dramatically altered in IFN-gamma knockout mice, demonstrating that IFN-gamma did not suppress the proliferation or differentiation of Th2 cells during MHV-68 infection. These observations indicate that IFN-gamma plays a nonessential or redundant role in the control of acute infection with MHV-68.

Modification of the Sendai virus-specific antibody and CD8+ T-cell responses in mice homozygous for disruption of the interleukin-4 gene

Homozygous disruption (-/-) of the interleukin-4 (IL-4) gene did not obviously modify the severity of Sendai virus infection in the highly susceptible 129/J mouse strain. The virus was cleared from the respiratory tract, and potent cytotoxic T lymphocyte (CTL) effectors were present in the cell population recovered by bronchoalveolar lavage. However, the prevalence of virus-specific CTL precursors (p) was consistently diminished in the spleen and regional lymph nodes of the IL-4 -/- mice at day 7 after infection. Also, virus-specific serum immunoglobulin G1 (IgG1) levels were greatly reduced and few IgG1-producing cells were detected in the lymphoid tissue. The effect on IgG1 class switching was to be expected, but the decrease in CTLp numbers has not been observed previously for a virus-specific immune response.

The cytotoxic T-lymphocyte response to Sendai virus is unimpaired in the absence of gamma interferon

Sendai virus is eliminated from the respiratory tract of gamma interferon (IFN-gamma) -/- BALB/c mice with normal kinetics. The level of virus-specific cytotoxic T-lymphocyte (CTL) activity in the cell population recovered by bronchoalveolar lavage is unimpaired, the prevalence of interleukin-4 (IL-4)-producing cells is increased, and the titers of virus-specific immunoglobulins IgG1 and IgG2b are higher in the IFN-gamma -/- mice. The emergence of this T-helper 2 response profile in both lymphoid tissue and the pneumonic lung has no obvious deleterious consequences. Virus clearance is slightly delayed following depletion of the CD4+ subset, with the effect being similar in magnitude for IFN-gamma -/- and +/+ mice. However, the generation of CTL precursors (CTLp) is diminished in the IFN-gamma -/- (but not +/+) mice in the absence of concurrent CD4+ T help. Apparently the clonal expansion of the CTLp population can be promoted either by a cytokine (perhaps IL-2) produced by the IFN-gamma -/- CD4+ T cells or by IFN-gamma made by other cell types in the +/+ mice.

Recombinant polyepitope vaccines for the delivery of multiple CD8 cytotoxic T cell epitopes

Development of epitope-based CD8 alpha beta CTL vaccines requires effective strategies for codelivery of large numbers of individual epitopes. We have designed an artificial "polyepitope" protein containing 10 contiguous minimal CTL epitopes, which were restricted by five MHC alleles and derived from five viruses, a parasite, and a tumor model. A recombinant vaccinia virus coding for this protein was capable of inducing MHC-restricted primary CTL responses to all 10 epitopes. Mice immunized with this recombinant vaccinia showed protection against murine cytomegalovirus, Sendai virus, and a tumor model. This simple generic approach to multiepitope delivery should find application in CTL-based vaccine design.

Recombinant polyepitope vaccines for the delivery of multiple CD8 cytotoxic T cell epitopes

Development of epitope-based CD8 alpha beta CTL vaccines requires effective strategies for codelivery of large numbers of individual epitopes. We have designed an artificial "polyepitope" protein containing 10 contiguous minimal CTL epitopes, which were restricted by five MHC alleles and derived from five viruses, a parasite, and a tumor model. A recombinant vaccinia virus coding for this protein was capable of inducing MHC-restricted primary CTL responses to all 10 epitopes. Mice immunized with this recombinant vaccinia showed protection against murine cytomegalovirus, Sendai virus, and a tumor model. This simple generic approach to multiepitope delivery should find application in CTL-based vaccine design.

Differential antigen burden modulates the gamma interferon but not the immunoglobulin response in mice that vary in susceptibility to Sendai virus pneumonia

Sendai virus, a paramyxovirus which causes murine pneumonia, grew to approximately 10-fold higher titers and was cleared less rapidly from the lungs of 129/J (129) than H-2b-compatible C57BL/6J (B6) mice. The more susceptible 129 mice also made higher titers of gamma interferon (IFN-gamma) and immunoglobulin G2a (IgG2a) virus-specific antibody. Analysis with acutely irradiated (950 rads) mice and immunologically reconstituted bone marrow (BM) radiation chimeras indicated that the enhanced virus growth was a function of the radiation-resistant respiratory epithelium. Prolonged exposure to more virus in turn influenced the magnitude of IFN-gamma production, most of which was made by CD4+ T lymphocytes. Somewhat surprisingly, however, the 129 pattern of a higher virus-specific serum Ig response skewed towards IgG2a mapped to the reconstituting BM. Thus, the characteristics of the humoral response are at least partly dissociated from both the antigen load, resulting from viral replication, and the level of IFN-gamma production. Further analysis of double chimeras (B6+129 BM-->B6 recipients) confirmed that the divergent humoral immune response to Sendai virus in B6 and 129 mice is largely determined by the inherent characteristics of the lymphoid cells.

Host response to Sendai virus in mice lacking class II major histocompatibility complex glycoproteins

The development of Sendai virus-specific cytotoxic T-lymphocyte (CTL) effectors and precursors (CTLp) has been compared for mice that are homozygous (-/-) for a disruption of the H-2I-Ab class II major histocompatibility complex glycoprotein and for normal (+/+) controls. The generation of CD8+ CTLp was not diminished in the -/- mice, though they failed to make virus-specific immunoglobulin G class antibodies. While the cellularity of the regional lymph nodes was decreased, the inflammatory process assayed by bronchoalveolar lavage (BAL) of the pneumonic lung was not modified, and potent CTL effectors were present in BAL populations recovered from both groups at day 10 after infection. There was little effect on virus clearance. Production of interleukin-2 by both freshly isolated BAL inflammatory cells and cultured lymph node cells was greatly diminished, though the -/- mice still made substantial levels of gamma interferon. However, treating the mice with a single dose of a monoclonal antibody to this cytokine, at least some of which is made by CD8+ T cells, did not decrease CTLp frequencies. As found previously with CD4-depleted H-2b mice, the development of Sendai virus-specific CD8+ T-cell-mediated immunity is not compromised by the absence of a concurrent class II major histocompatibility complex-restricted response.

Induction of cytokines in mice with parainfluenza pneumonia

The possible involvement of cytokines in the acute viral pneumonia induced by the murine parainfluenza type 1 virus, Sendai virus, was studied. Cytokine profiles for both the respiratory tract and the draining mediastinal lymph node (MLN) of virus-infected C57BL/6J mice were quantified by using the single-cell cytokine (ELISPOT) assay with freshly isolated cell populations and enzyme-linked immunosorbent assay for lung lavage fluids and culture supernatants. Maximal levels of interleukin 2 (IL-2), gamma interferon (IFN-gamma), tumor necrosis factor, IL-6, and IL-10 were detected at the inflammatory site 7 to 10 days after infection, about the time that virus is cleared from the lung. The frequencies of cells producing IL-2, IL-4, IL-6, IL-10, IFN-gamma, and tumor necrosis factor were much higher for the bronchoalveolar lavage (BAL) cell population than for the MLN cell population. Cytokine production after in vitro restimulation of MLN cells was dominated by IL-2 and IFN-gamma, with low levels of IL-10 and IL-6 also being present. Most of the cytokine was produced by the CD4+ cells, although the CD8+ subset was also involved. No IL-4 was found in the BAL fluid or in culture supernatants from restimulated BAL or MLN cells, although a high frequency of IL-4-producing cells was demonstrated in the BAL population by ELISPOT analysis.