iNKT cells prevent obesity-induced hepatic steatosis in mice in a C-C chemokine receptor 7-dependent manner.

Non-alcoholic fatty liver disease and non-alcoholic steatohepatitis are characterized by an increase in hepatic triglyceride content with infiltration of immune cells, which can cause steatohepatitis and hepatic insulin resistance. C-C chemokine receptor 7 (CCR7) is primarily expressed in immune cells, and CCR7 deficiency leads to the development of multi-organ autoimmunity, chronic renal disease and autoimmune diabetes. Here, we investigated the effect of CCR7 on hepatic steatosis in a mouse model and its underlying mechanism. Our results demonstrated that body and liver weights were higher in the CCR7-/- mice than in the wild-type (WT) mice when they were fed a high-fat diet. Further, glucose tolerance and insulin sensitivity were markedly diminished in CCR7-/- mice. The number of invariant natural killer T (iNKT) cells was reduced in the livers of the CCR7-/- mice. Moreover, liver inflammation was detected in obese CCR7-/- mice, which was ameliorated by the adoptive transfer of hepatic mononuclear cells from WT mice, but not through the transfer of hepatic mononuclear cells from CD1d-/- or interleukin-10-deficient (IL-10-/-) mice. Overall, these results suggest that CCR7+ mononuclear cells in the liver could regulate obesity-induced hepatic steatosis via induction of IL-10-expressing iNKT cells.

Gut commensal Bacteroides acidifaciens prevents obesity and improves insulin sensitivity in mice.

In humans, the composition of gut commensal bacteria is closely correlated with obesity. The bacteria modulate metabolites and influence host immunity. In this study, we attempted to determine whether there is a direct correlation between specific commensal bacteria and host metabolism. As mice aged, we found significantly reduced body weight and fat mass in Atg7ΔCD11c mice when compared with Atg7f/f mice. When mice shared commensal bacteria by co-housing or feces transfer experiments, body weight and fat mass were similar in both mouse groups. By pyrosequencing analysis, Bacteroides acidifaciens (BA) was significantly increased in feces of Atg7ΔCD11c mice compared with those of control Atg7f/f mice. Wild-type C57BL/6 (B6) mice fed with BA were significantly more likely to gain less weight and fat mass than mice fed with PBS. Of note, the expression level of peroxisome proliferator-activated receptor alpha (PPARα) was consistently increased in the adipose tissues of Atg7ΔCD11c mice, B6 mice transferred with fecal microbiota of Atg7ΔCD11c mice, and BA-fed B6 mice. Furthermore, B6 mice fed with BA showed elevated insulin levels in serum, accompanied by increased serum glucagon-like peptide-1 and decreased intestinal dipeptidyl peptidase-4. These finding suggest that BA may have potential for treatment of metabolic diseases such as diabetes and obesity.

Immunosuppressive properties of human amniotic membrane for mixed lymphocyte reaction.

The combination of allograft limbal transplantation (ALT) and amniotic membrane transplantation (AMT) has been applied in the treatment of severe ocular surface diseases. The beneficial effect of this combination has been thought to result from possible immunosuppressive ability of amniotic membrane (AM). However, the mechanisms of any such ability remain unknown. In this study, we investigated whether human AM has the ability to suppress allo-reactive T cell responses in vitro. For mixed lymphocyte reaction (MLR), lymphocytes isolated from lymph nodes of C57BL/6 mice (Mls1b, Vbeta6+) were cultured with irradiated splenocytes from DBA/2 mice (Mls1a, Vbeta6-) with or without human AM. For carboxyfluorescein diacetate succinimidyl ester (CFSE) experiments, responder lymph node cells were labelled with a stable intracellular fluorescent dye and cultured with irradiated stimulator cells. The ratio of responder Vbeta6+ T cells was then determined by FACS analysis, and the division profiles of responder Vbeta6+ T cells were analysed by CFSE content. Furthermore, Th1 and Th2 cytokine synthesis by allo-reactive T cells in MLR culture supernatants was determined by enzyme-linked immunosorbent assay (ELISA). Addition of AM to the MLR culture resulted in the significant inhibition of thymidine incorporation compared with control culture lacking AM. The population of responder CD4+Vbeta6+ T cells was significantly reduced in the AM-treated culture in comparison to control. CFSE analysis revealed less division and lower proliferation of responder CD4+Vbeta6+ T cells in cultures with AM than without. In addition, allo-rective T cell synthesis of both Th1 (IL-2 and IFNgamma) and Th2 (IL-6 and IL-10) type cytokine was significantly decreased in the presence of AM. These results indicate that human AM has the ability to suppress allo-reactive T cells in vitro. This inhibitory effect likely contributes to the success of the ALT-AMT combination.

Development of antigen induced colitis in SCID mice reconstituted with spleen derived memory type CD4(+) CD45RB(+) T cells.

BACKGROUND AND AIMS: Enteric bacterial and/or food antigens may be crucial in the development of colitis but little is known of the exact mechanism of antigen specific reactions in this condition. The aim of this study was to determine whether systemically primed antigen specific CD4(+) T cells containing both CD45RB(high) and CD45RB(low) populations participate as a pathogenic subset that in turn leads to inflammatory reactions selectively in the large intestine. METHODS: SCID mice were reconstituted with splenic CD4(+) CD45RB(+) T cells or CD4(+) CD45RB(low) T cells isolated from donor mice systemically primed with ovalbumin (OVA) plus CFA. The reconstituted mice were then fed OVA for several weeks. RESULTS: Reconstitution of SCID mice with OVA primed splenic CD4(+) T cells, containing populations of CD45RB(high) and CD45RB(low), resulted in the development of colitis by 4-5 weeks following repeated administration of oral OVA. Histopathological study revealed thickened wall, inflammatory cell infiltration, crypt elongation, and loss of goblet cells in the large intestine. The CD4(+) CD45RB(low) population of cells extracted from the affected large intestine secreted high levels of interferon gamma (IFN-gamma) and tumour necrosis factor alpha (TNF-alpha) at the protein and mRNA levels. Administration of neutralising antibodies to TNF-alpha, but not to IFN-gamma, prevented the development of colitis. Furthermore, adoptive transfer with OVA primed splenic CD4(+) CD45RB(low) T cells evoked severe colitis. CONCLUSIONS: These results demonstrate that systemically primed activated/memory CD4(+) CD45RB(low) T cells can mediate the development of specific antigen induced colitis in SCID mice, and also that TNF-alpha is critical in the induction of this type of colitis. Our results contrast with those from studies in some colitis models in which CD45RB(low) T cells appeared to prevent colitis through secretion of immunosuppressive cytokines.

Nasal administration of cholera toxin (CT) suppresses clinical signs of experimental autoimmune encephalomyelitis (EAE).

Cholera toxin (CT), a major enterotoxin produced by Vibrio cholerae, elicits mucosal adjuvant activities by inducing antigen-specific CD4+ T cells secreting T helper type 2 (Th2) cytokines. Experimental autoimmune encephalomyelitis (EAE) is induced by Th1 cells specific for myelin-derived antigens. We induced EAE in C57BL/6 mice with myelin oligodendrocyte glycoprotein (MOG) 35-55 and CT was nasally administered as an immunomodulator on day 7 following MOG challenge. Clinical severity in the CT-treated mice was milder when compared to PBS-treated mice, while the levels of expression of interleukin (IL)-12 and interferon (IFN)-gamma in the central nervous system (CNS) of CT-treated mice were lower than PBS-treated mice. Thus, nasal administration of the mucosal immunomodulator CT ameliorated the severity of EAE, which was associated with the suppression of Th1 cell responses.

The mucosal immune system: from specialized immune defense to inflammation and allergy.

The mucosal immune system is a first line of defense against foreign antigens, including microbial and dietary antigens. Under normal circumstances, the mucosal immune system employs tightly regulated dynamic mucosal intra- and internets consisting of inductive and effector sites for the induction of an appropriate immunological homeostasis between the host and mucosal environments. The common mucosal immune system (CMIS), which interconnects between inductive (e.g. Peyer patch) and effector (e.g. intestinal lamina propria) tissues for the induction of the IgA response, is well characterized. Recent results provide strong evidence for the presence of a CMIS-independent IgA induction pathway. Two distinct subsets of mucosal IgA-committed B cells termed B-1 and B-2, are associated with CMIS-independence and CMIS-dependent cascades respectively. In some cases, the breakdown of this tightly regulated mucosal immune system leads to pathological responses to different gut environmental antigens. As a result, disorders such as inflammatory bowel disease (e.g. IBD) and allergic gastroenteropathy can be evoked in the gastrointestinal tissues. Recently, many studies have described possible molecular and cellular mechanisms for this dysfunction in the gastrointestinal tissues by using murine models with specific gene manipulation. In this review we summarize recent findings from our group concerning the CMIS-dependent and CMIS-independent IgA induction pathways and gastrointestinal diseases (IBD and intestinal allergic diseases). These observations may provide useful information for the development of new mucosal immune therapy.

Systemically derived large intestinal CD4(+) Th2 cells play a central role in STAT6-mediated allergic diarrhea.

Systemically primed BALB/c mice developed severe diarrhea after repeated oral administration of ovalbumin (OVA). Histological analysis demonstrated that dramatic infiltration of eosinophils and mast cells occurred in the large intestine but not in the small intestine of mice with diarrhea. Interestingly, CD4(+) alphabeta T cells of the large intestine secreted IL-4 and IL-13 at high levels. Identically treated STAT6 gene-disrupted mice failed to develop OVA-induced diarrhea. Further, treatment of BALB/c mice with monoclonal anti-IL-4 antibody prevented the development of allergic diarrhea. An adoptive transfer study showed that systemically primed splenic CD4(+) T cells were preferentially recruited into the large intestine upon exposure to oral OVA. These results strongly suggest that systemically derived CD4(+) alphabeta T cells of the large intestine play a critical role in the onset of Th2-mediated intestinal allergic disorders via STAT6 signal transduction.

Enterotoxin adjuvants have direct effects on T cells and antigen-presenting cells that result in either interleukin-4-dependent or -independent immune responses.

In an in vitro study, Escherichia coli heat-labile toxin (LT) was shown to directly affect activated CD4(+) T cells and support interleukin (IL)-5 production in IL-4-deficient (IL-4(-/-)) mice, whereas cholera toxin (CT) did not. Both LT and CT enhanced B7-2 expression on B cells and macrophages. These effects were not influenced by CD40-CD40 ligand cosignaling. Addition of LT- or CT-treated antigen-presenting cells to anti-CD3-triggered CD4(+) T cells resulted in the induction of T cell proliferative responses. Further, these responses were inhibited by anti-B7-2 monoclonal antibody. Cocultivation of CD4(+) T cells with LT- or CT-treated antigen-presenting cells and anti-CD3 enhanced Th1- and IL-4-mediated Th2-type cytokine production. The results from in vitro studies were supported by in vivo studies in IL-4(-/-) mice, in which LT induced mucosal IgA responses but CT did not. Thus, although both LT and CT induce mucosal adjuvant responses via IL-4-dependent Th2-type responses, LT also elicits Th1- and IL-4-independent Th2-type responses.

Alternate mucosal immune system: organized Peyer's patches are not required for IgA responses in the gastrointestinal tract.

The progeny of mice treated with lymphotoxin (LT)-beta receptor (LTbetaR) and Ig (LTbetaR-Ig) lack Peyer's patches but not mesenteric lymph nodes (MLN). In this study, we used this approach to determine the importance of Peyer's patches for induction of mucosal IgA Ab responses in the murine gastrointestinal tract. Immunohistochemical analysis revealed that LTbetaR-Ig-treated, Peyer's patch null (PP null) mice possessed significant numbers of IgA-positive (IgA+) plasma cells in the intestinal lamina propria. Further, oral immunization of PP null mice with OVA plus cholera toxin as mucosal adjuvant resulted in Ag-specific mucosal IgA and serum IgG Ab responses. OVA-specific CD4+ T cells of the Th2 type were induced in MLN and spleen of PP null mice. In contrast, when TNF and LT-alpha double knockout (TNF/LT-alpha-/-) mice, which lack both Peyer's patches and MLN, were orally immunized with OVA plus cholera toxin, neither mucosal IgA nor serum IgG anti-OVA Abs were induced. On the other hand, LTbetaR-Ig- and TNF receptor 55-Ig-treated normal adult mice elicited OVA- and cholera toxin B subunit-specific mucosal IgA responses, indicating that both LT-alphabeta and TNF/LT-alpha pathways do not contribute for class switching for IgA Ab responses. These results show that the MLN plays a more important role than had been appreciated until now for the induction of both mucosal and systemic Ab responses after oral immunization. Further, organized Peyer's patches are not a strict requirement for induction of mucosal IgA Ab responses in the gastrointestinal tract.

Comparison of nasal and oral tolerance for the prevention of collagen induced murine arthritis.

OBJECTIVE: Administration of bovine type II collagen (CII) or of its peptide either orally or nasally has been reported to suppress the development of collagen induced arthritis (CIA) in mice and rats. We examined the inhibitory effects of CII delivered by each route on CIA in DBA/1J mice to determine which route was superior. METHODS: Male mice were injected twice with CII in Freund's complete adjuvant to induce CIA. Before induction of CIA, 1, 10, or 40 microg of CII were administered nasally 15 times and 10, 100, 500, or 1000 microg of CII were given 10 times orally. The development of arthritis, arthritis score, CII-specific delayed-type hypersensitivity (DTH) response, and CII-specific antibody levels were examined. RESULTS: Nasal administration of 10 microg of CII 15 times had the most prominent suppressive effects, reducing disease incidence by 50% and inhibiting both CII-specific IgG antibody and DTH responses. Of all the mice undergoing oral administration, those receiving 500 microg of CII 10 times showed the greatest suppressive potential. However, the treatment only delayed disease onset for roughly 3 weeks, lowering CII-specific IgG antibody levels but failing to suppress DTH responses. CONCLUSION: Nasal administration of CII reduced CIA development and inhibited CII-specific T cell and antibody responses to a greater degree than did oral administration.

gammadelta T cells regulate mucosally induced tolerance in a dose-dependent fashion.

We used gammadelta TCR-deficient (TCRdelta(-/-)) mice to examine the role of gammadelta T cells for induction of mucosal responses and systemic tolerance to high versus low doses of oral antigen. When either TCRdelta(-/-) or TCRdelta(+/+) mice were immunized orally with a high dose of ovalbumin (OVA) prior to parenteral challenge, systemic IgG and IgE antibody responses were markedly reduced in both types of mice, while mucosal IgA responses were reduced only in the TCRdelta(-/-) mice. Reduced T cell proliferative responses and delayed-type hypersensitivity were seen in TCRdelta(-/-) and TCRdelta(+/+) mice given the high dose of OVA. Antigen-induced T(h)1 and T(h)2 cytokine production by splenic CD4(+) T cells was severely inhibited in orally tolerized TCRdelta(-/-) and TCRdelta(+/+) mice. In contrast, while oral tolerance associated with increased levels of IL-10 synthesis was induced by a low dose of OVA in TCRdelta(+/+) mice, the TCRdelta(-/-) mice were not tolerized and failed to produce IL-10. Our findings indicate that gammadelta T cells play a significant immunoregulatory role in IL-10-mediated, low-dose oral tolerance induction, but are not essential participants in the induction of systemic tolerance to orally introduced antigens given in larger doses.

Partial IgA-deficiency with increased Th2-type cytokines in TGF-beta 1 knockout mice.

Though it has been shown that TGF-beta 1 directs B cells to switch to IgA in vitro, no studies have assessed TGF-beta 1 effects on mucosal vs systemic immunity in vivo. When the B cell functions of TGF-beta 1 gene-disrupted (TGF-beta 1-/-) mice were analyzed, significantly decreased IgA levels and increased IgG and IgM levels in serum and external secretions were observed. Further, analysis of Ab forming cells (AFC) isolated from both mucosal and systemic lymphoid tissue showed elevated IgM, IgG, and IgE, with decreased IgA AFC. A lack of IgA-committed B cells was seen in TGF-beta 1-/- mice, especially in the gastrointestinal (GI) tract. Splenic T cells triggered via the TCR expressed elevated Th2-type cytokines and, consistent with this observation, a 31-fold increase in serum IgE was seen in TGF-beta 1-/- mice. Thus, uncontrolled B cell responses, which include elevated IgE levels, a lack of antiinflammatory IgA, and an excess of complement-binding IgG and IgM Abs, will promote inflammation at mucosal surfaces in TGF-beta 1-/- mice and likely contribute to pulmonary and GI tract lesions, ultimately leading to the early death of these mice.

Direct effects on antigen-presenting cells and T lymphocytes explain the adjuvanticity of a nontoxic cholera toxin mutant.

The present study has elucidated two distinct mechanisms that may explain how a mutant of cholera toxin (mCT), E112K, retains adjuvant effects though it lacks ADP-ribosyltransferase activity and associated toxicity. In the first mechanism, we show that mCT E112K, like native cholera toxin (nCT), enhances B7-2 expression, but, to some extent, also enhances B7-1 on Peyer's patch B cells and macrophages. Cocultivation of CD4+ T cells with E112K- or nCT-treated B cells and macrophages in the presence of anti-CD3 stimulation resulted in the induction of T cell-proliferative responses. Further, the responses were blocked by mAbs to B7-1 and/or B7-2; however, the effect of anti-B7-1 was minimal. In the second mechanism, addition of mCT E112K or nCT to anti-CD3 mAb-stimulated Peyer's patch CD4+ T cells inhibited proliferative responses, while recombinant CT-B subunit (rCT-B) did not. Analysis of cytokine responses showed that both mCT E112K and nCT preferentially inhibited IFN-gamma production. Interestingly, however, nCT, but not mCT E112K, induced apoptosis in CD4+ T cells activated via the TCR-CD3 complex. These results indicate that CT uses at least two pathways for inhibition of Th1 responses and that, while nCT induces cAMP accumulation that in turn leads to apoptosis in Th1-type cells, mCT E112K, which lacks ADP-ribosyltransferase activity, inhibits IFN-gamma synthesis by a separate mechanism. Thus, mCT E112K, like nCT, induces adjuvant responses via up-regulation of mainly B7-2 on APCs and through preferential inhibition of Th1-type CD4+ T cell responses in the absence of ADP-ribosyltransferase activity.

Mucosally induced systemic T cell unresponsiveness to ovalbumin requires CD40 ligand-CD40 interactions.

CD40 ligand (CD40L) gene-disrupted (CD40L-/-) mice were employed to examine the role of costimulatory signals via CD40L-CD40 interactions in mucosally induced tolerance. CD40L-/- and control (CD40L+/+) mice of the same C57BL/6 x 129/J background were immunized orally with 25 mg of OVA before systemic challenge with OVA in CFA. While CD40L+/+ mice showed reductions in Ag-specific T cell responses including delayed-type hypersensitivity (DTH) and proliferative responses, CD40L-/- mice underwent normal T cell responses. Further, cytokine analysis of splenic CD4+ T cells showed that both Th1-type (e.g., IFN-gamma and IL-2) and Th2-type (e.g., IL-4, IL-5, IL-6, and IL-10) responses were maintained in CD40L-/- mice orally immunized with OVA, whereas these cytokine responses in CD40L+/+ mice were significantly reduced. In addition, splenic CD4+ T cells from CD40L-/- mice orally immunized with OVA provided B cell help in Ag-specific Ab-forming cells when the cells were cultured with naive B cells in the presence of Ag and CD40L-transfected cell lines. In contrast, an identical culture condition containing splenic CD4+ T cells from orally tolerized CD40L+/+ mice did not exhibit helper activity. Taken together, these findings indicate that CD40L and CD40 interactions are essential for the induction of systemic T cell unresponsiveness to orally administered Ag.

Lack of orally induced systemic unresponsiveness in IFN-gamma knockout mice.

Splenic T cells isolated from BALB/c mice that had been mucosally tolerized by oral administration of 25 mg of OVA revealed selective increases in IFN-gamma production with impaired levels of IL-2, IL-4, IL-5, and IL-10. These mice possessed reduced splenic OVA-specific T cell proliferative and delayed-type hypersensitivity responses when compared with nontolerized controls. Further, OVA-specific IgG Ab responses in serum and the numbers of IgG Ab-forming cells in spleen were significantly diminished following systemic challenge with OVA in CFA. When IFN-gamma-deficient (IFN-gamma-/-) mice of the same genetic background were given an oral dose of 25 mg of OVA before systemic immunization, no reduction in OVA-specific IgG Ab responses in serum and spleen was seen. Furthermore, the serum IgG Ab responses were restricted to IgG1 and IgG2b subclasses. Interestingly, although IFN-gamma-/- mice displayed a partial diminution of T cell proliferative and delayed-type hypersensitivity responses to OVA, significant responses were still present when compared with the low responses noted in IFN-gamma+/+ mice. In addition, OVA-specific T cells from IFN-gamma-/- mice produced Th2-type cytokines (e.g., IL-4), which provided help for systemic OVA-specific serum IgG1 and IgG2b Ab responses. These findings clearly indicate a central role for IFN-gamma in the induction and maintenance of mucosally induced tolerance.

[The mechanisms of oral tolerance].

Oral or nasal administration of a single high dose or repeated mucosal delivery of low doses of proteins have been shown to induce systemic unresponsiveness in the presence of mucosal IgA responses. The induction of oral tolerance(or mucosally-induced tolerance) is mediated by T cells involved in the generation of active suppression, clonal anergy or clonal deletion. Studies of T helper(Th) cytokine responses have suggested that Th1- and Th2-type cells are involved in the induction of oral tolerance. Further, gamma delta T cells appear to be an important T cell subset for the regulation of oral tolerance.

gamma/delta T cell-deficient mice have impaired mucosal immunoglobulin A responses.

Mucosal tissues of mice are enriched in T cells that express the gamma/delta T cell receptor. Since the function of these cells remains unclear, we have compared mucosal immune responses in gamma/delta T cell receptor-deficient (TCRdelta-/-) mice versus control mice of the same genetic background. The frequency of intestinal immunoglobulin (Ig) A plasma cells as well as IgA levels in serum, bile, saliva, and fecal samples were markedly reduced in TCRdelta-/- mice. The TCRdelta-/- mice produced much lower levels of IgA antibodies when immunized orally with a vaccine of tetanus toxoid plus cholera toxin as adjuvant. Conversely, the antigen-specific IgM and IgG antibody responses were comparable to orally immunized control mice. Direct assessment of the cells forming antibodies against the tetanus toxoid and cholera toxin antigens indicated that significantly lower numbers of IgA antibody-producing cells were present in the intestinal lamina propria and Peyer's patches of TCRdelta-/- mice compared with the orally immunized control mice. The selective reduction of IgA responses to ingested antigens in the absence of gamma/delta T cells suggests a specialized role for gamma/delta cells in mucosal immunity.

Ultracytochemistry of thymic epithelial cells in rats given organotin.

A combined light-and electron microscopic study of thymic changes in the adult rats after di-n-butyltin dichloride (DBTC) administration has been made. A rapid depletion of thymocytes in the cortex of thymus and subsequent rapid recovery of the number of thymocytes occurred. In this process, at 1-3 days after DBTC treatment the number of necrotic thymocytes was maximal. At 3-6 days, reticular epithelial cells were predominantly phagocytizing and acid phosphatase-positive in the cortex and cortico-medullary regions, where they appeared to develop from macrophage-like cells. During acute involution, it was likely that reticular epithelial cells were phagocytic and remove the necrotic thymocytes. The proportion of CD4+ CD8+ cells in the cortex of thymus was maximally reduced from day 3 onwards and reached the lowest level at 6 days after single oral dose of DBTC. On these days, the proportions of CD4- CD8- and single positive cell (CD4- CD8+ or CD4+ CD8-) subsets were relatively increased. These data suggest that DBTC preferentially causes an initial depletion of CD4+ CD8+ cells in thymus, and both macrophages and reticular epithelial cells of the cortex may be involved in the rapid removal of damaged thymocytes from thymus.

A non-glycosylated form of pulmonary surfactant protein A appears in rat amniotic fluid.

Surfactant protein A (SP-A) is a family of glycoproteins that have a triplet with 26, 32 and 36 kDa under reducing conditions in rat lung. We wanted to evaluate the SP-A forms in amniotic fluid of pregnant rats compared to those found in rat lungs. By Western blot analysis, glycosylated SP-A, was not found in the amniotic fluid in contrast to the pulmonary surfactant triplet SP-A, which comprises a 26 kDa protein and its glycosylated 32 and 36 kDa forms. The SP-A concentration in amniotic fluid was barely detectable at 18 days of gestation (20 +/- 12 ng.ml-1), and then increased and reached 700 +/- 333 ng.ml-1 at the final gestational day 21, as determined by an enzyme-linked immunoabsorbent assay. Immunohistochemically, SP-A was found in some epithelial cells of larger respiratory bronchi, but not, or to a lesser degree, in smaller respiratory bronchi at gestational day 18. At 21 days of gestation, SP-A was detected in bronchial and bronchiolar nonciliated epithelial Clara cells, alveolar epithelial type II cells and some alveolar macrophages. The ratio of the 26, 32 and 36 kDa SP-A forms in bronchoalveolar, bronchobronchiolar and tracheal lavage fluids prepared from adult rats was 6:29:65, 84:5:11 and 100:0:0, respectively. These findings show the presence of a non-glycosylated SP-A in rat amniotic fluid. This may reflect the increased ratio of non-glycosylated SP-A to bronchoalveolar, bronchobronchiolar and tracheal lavage fluids, respectively.