Characterization of human sialoadhesin, a sialic acid binding receptor expressed by resident and inflammatory macrophage populations.

Sialoadhesin is a macrophage-restricted cellular interaction molecule and a prototypic member of the Siglec family of sialic acid binding immunoglobulin (Ig)-like lectins. So far, it has only been characterized in rodents. Here, we report the molecular cloning, binding properties, and expression pattern of human sialoadhesin. The predicted protein sequences of human and mouse sialoadhesin are about 72% identical, with the greatest similarity in the extracellular region, which comprises 17 Ig domains in both species. A recombinant protein consisting of the first 4 N-terminal domains of human sialoadhesin fused to the Fc region of human IgG1 mediated sialic acid-dependent binding with a specificity similar to its mouse counterpart, preferring sialic acid in the alpha2,3 glycosidic linkage over the alpha2,6 linkage. By flow cytometry with peripheral blood leukocytes, recombinant sialoadhesin bound strongly to granulocytes with intermediate binding to monocytes, natural killer cells, B cells, and a subset of CD8 T cells. Using antibodies raised to the recombinant protein, sialoadhesin was immunoprecipitated from the THP-1 human monocytic cell line as an approximate 200-kd glycoprotein. The expression pattern of human sialoadhesin was found to be similar to that of the mouse receptor, being absent from monocytes and other peripheral blood leukocytes, but expressed strongly by tissue macrophages in the spleen, lymph node, bone marrow, liver, colon, and lungs. High expression was also found on inflammatory macrophages present in affected tissues from patients with rheumatoid arthritis.

Flow cytometry : measurement of leukocyte receptor expression and function.

Flow cytometry is an invaluable tool for the analysis of leukocyte populations in inflammation. Flow cytometers, particularly those designed purely for analysis rather than cell sorting, have now become user-friendly machines and are commonplace in many laboratories. Any cell type can be analyzed, as long as a single-cell suspension can be prepared, and small numbers of cells can be used, often without the need for purification, thus making measurements rapid and less susceptible to artifacts associated with lengthy cell separation procedures.

Basophil responses to chemokines are regulated by both sequential and cooperative receptor signaling.

To investigate human basophil responses to chemokines, we have developed a sensitive assay that uses flow cytometry to measure leukocyte shape change as a marker of cell responsiveness. PBMC were isolated from the blood of volunteers. Basophils were identified as a single population of cells that stained positive for IL-3Ralpha (CDw123) and negative for HLA-DR, and their increase in forward scatter (as a result of cell shape change) in response to chemokines was measured. Shape change responses of basophils to chemokines were highly reproducible, with a rank order of potency: monocyte chemoattractant protein (MCP) 4 (peak at <1 nM) >/= eotaxin-2 = eotaxin-3 >/= eotaxin > MCP-1 = MCP-3 > macrophage-inflammatory protein-1alpha > RANTES = MCP-2 = IL-8. The CCR4-selective ligand macrophage-derived chemokine did not elicit a response at concentrations up to 10 nM. Blocking mAbs to CCR2 and CCR3 demonstrated that responses to higher concentrations (>10 nM) of MCP-1 were mediated by CCR3 rather than CCR2, whereas MCP-4 exhibited a biphasic response consistent with sequential activation of CCR3 at lower concentrations and CCR2 at 10 nM MCP-4 and above. In contrast, responses to MCP-3 were blocked only in the presence of both mAbs, but not after pretreatment with either anti-CCR2 or anti-CCR3 mAb alone. These patterns of receptor usage were different from those seen for eosinophils and monocytes. We suggest that cooperation between CCRs might be a mechanism for preferential recruitment of basophils, as occurs in tissue hypersensitivity responses in vivo.

Eotaxin and RANTES expression by the dermal endothelium is associated with eosinophil infiltration after ivermectin treatment of onchocerciasis.

The roles of eotaxin, RANTES, and MCP-3 expression in eosinophil recruitment to the site of parasite killing that occurs following ivermectin treatment of onchocerciasis were assessed in the skin of 13 Onchocerca volvulus-infected subjects and two noninfected controls before and after ivermectin treatment. Adverse reactions in infected subjects were associated with the appearance of eosinophils in the dermis as part of a perivascular inflammatory infiltrate. Although no expression of RANTES and eotaxin was seen in dermal vascular endothelial cells in biopsies taken before treatment (nor at any time in the skin of uninfected controls), endothelial expression of both eotaxin and RANTES was noted by 24 h following treatment. While RANTES expression was transient, eotaxin expression increased in parallel with increasing eosinophil recruitment up to 60 h posttreatment. These observations indicate that endothelial expression of eotaxin and RANTES may have an important role in eosinophil recruitment into the skin during helminth-killing reactions.

[The role of chemokines in nasal polyps]. / Die Rolle von Chemokinen bei Nasenpolypen.

Nasal polyposis is an inflammatory condition of the nose and the sinuses characterised by a marked infiltration of eosinophils in addition to lymphocytes, mast cells and macrophages. The selective recruitment of eosinophils to inflammatory sites is mediated by CC chemokines such as Eotaxin and Eotaxin-2. In the present study histology, immunohistochemistry and ELISA were performed. The levels of Eotaxin and Eotaxin-2 and for comparison other chemokines RANTES and IL-8 were measured in nasal polyp tissue and in control nasal tissue. On histological examination 6 polyps showed an oedematous structure, one was glandular and one had a fibromatous pattern, while all showed a marked eosinophil infiltration. Immunohistochemistry of the polyps showed that epithelial cells were strongly positive for Eotaxin and IL-8, whereas endothelial cells stained positive for Eotaxin-2. Significantly higher amounts of Eotaxin, Eotaxin-2 and IL-8 were detected in polyp tissue when compared with control middle turbinates. The increased levels of eosinophil-stimulating chemokines, such as Eotaxin and Eotaxin-2 in nasal polyps suggest that they may be important regulators of eosinophil recruitment in this inflammatory disease.

Macrophage-tumour cell interactions: identification of MUC1 on breast cancer cells as a potential counter-receptor for the macrophage-restricted receptor, sialoadhesin.

In many carcinomas, infiltrating macrophages are commonly found closely associated with tumour cells but little is known concerning the nature or significance of adhesion molecules involved in these cellular interactions. Here we demonstrate in primary human breast cancers that sialoadhesin (Sn), a macrophage-restricted adhesion molecule, is frequently expressed on infiltrating cells that often make close contact with breast carcinoma cells. To determine whether Sn could act as a specific receptor for ligands on breast cancer cell lines, binding assays were performed with a recombinant form of the protein fused to the Fc portion of human immunoglobulin G1 (IgG1) (Sn-Fc). Sn-Fc was found to bind specifically and in a sialic acid-dependent manner to the breast cancer cell lines MCF-7, T47.D and BT-20 both in solid- and solution-phase binding assays. To investigate the nature of the sialoglycoproteins recognized by Sn on breast cancer cells, MCF-7 cells were labelled with [6-3H]glucosamine. Following precipitation with Sn-Fc, a major band of approximately 240000 MW was revealed, which was shown in reprecipitation and Western blotting experiments to be the epithelial mucin, MUC1.

Human eotaxin induces eosinophil extravasation through rat mesenteric venules: role of alpha4 integrins and vascular cell adhesion molecule-1.

Eotaxin is a potent eosinophil-specific CC-chemokine, which has been shown to play a role in the selective induction of eosinophil accumulation in a number of allergic models of inflammation. Many aspects of the mechanism by which eotaxin induces eosinophil accumulation in vivo remain unresolved. In the present study, we investigated the direct effect of synthetic human eotaxin on leucocyte/endothelial cell interactions within rat mesenteric venules, as quantified by intravital microscopy. Topical eotaxin (30 pmol) induced rapid firm adhesion and extravasation of leucocytes within the rat mesentery, the extravasated leucocytes all being eosinophils, as determined by histological analysis. Whilst eotaxin was unable to stimulate the interaction of rat eosinophils with vascular cell adhesion molecule-1 (VCAM-1) under static conditions in vitro, eotaxin-induced responses in vivo were significantly suppressed by anti-alpha4 integrin and anti-VCAM-1 monoclonal antibodies (mAbs). The anti-alpha4 integrin mAb, HP2/1 (3.5 mg/kg), inhibited the eotaxin-induced firm adhesion and extravasation, 60 min postapplication of the chemokine, by 89% and 84%, respectively. In the same set of experiments, the anti-VCAM-1 mAb, 5F10 (3.5 mg/kg), inhibited leucocyte adhesion and extravasation by 61% and 63%, respectively. These results demonstrate that eotaxin-induced migration of eosinophils through rat mesenteric venules in vivo is dependent on an alpha4 integrin/VCAM-1 adhesion pathway, the significance of which may only be evident under flow conditions and/or following the ligation of other adhesion molecules expressed on eosinophils.

Differential regulation of eosinophil chemokine signaling via CCR3 and non-CCR3 pathways.

To investigate eosinophil stimulation by chemokines we developed a sensitive assay of leukocyte shape change, the gated autofluorescence/forward scatter assay. Leukocyte shape change responses are mediated through rearrangements of the cellular cytoskeleton in a dynamic process typically resulting in a polarized cell and are essential to the processes of leukocyte migration from the microcirculation into sites of inflammation. We examined the actions of the chemokines eotaxin, eotaxin-2, monocyte chemoattractant protein-1 (MCP-1), MCP-3, MCP-4, RANTES, macrophage inflammatory protein-1alpha (MIP-1alpha), and IL-8 on leukocytes in mixed cell suspensions and focused on the responses of eosinophils to C-C chemokines. Those chemokines acting on CCR3 induced a rapid shape change in eosinophils from all donors; of these, eotaxin and eotaxin-2 were the most potent. Responses to MCP-4 were qualitatively different, showing marked reversal of shape change responses with agonist concentration and duration of treatment. In contrast, MIP-1alpha induced a potent response in eosinophils from a small and previously undescribed subgroup of donors via a non-CCR3 pathway likely to be CCR1 mediated. Incubation of leukocytes at 37 degrees C for 90 min in the absence of extracellular calcium up-regulated responses to MCP-4 and MIP-1alpha in the majority of donors, and there was a small increase in responses to eotaxin. MIP-1alpha responsiveness in vivo may therefore be a function of both CCR1 expression levels and the regulated efficiency of coupling to intracellular signaling pathways. The observed up-regulation of MIP-1alpha signaling via non-CCR3 pathways may play a role in eosinophil recruitment in inflammatory states such as occurs in the asthmatic lung.

Elevated concentrations of eotaxin and interleukin-5 in human neurocysticercosis.

Symptomatic neurocysticercosis, a major cause of epilepsy worldwide, results from inflammation around Taenia solium larvae, but the mechanisms are unknown. Eotaxin, not previously reported in cases of human infection, and interleukin-5 (IL-5) but not IL-8 concentrations were elevated in patient serum, and IL-5 levels were also elevated in cerebrospinal fluid (CSF). Eosinophil-selective mediators may be involved in the pathogenesis of cysticercosis. IL-6 concentrations were also elevated in patient CSF, possibly indicative of an acute-phase response.

Tumor necrosis factor alpha-induced eosinophil accumulation in rat skin is dependent on alpha4 integrin/vascular cell adhesion molecule-1 adhesion pathways.

Tumor necrosis factor alpha (TNFalpha) is a cytokine implicated in the pathogenesis of numerous chronic and acute inflammatory conditions. In the present study, we have characterized the ability of TNFalpha in inducing eosinophil accumulation in rat skin and have shown the inhibitory effects of anti-alpha4 integrin and anti-vascular cell adhesion molecule-1 (VCAM-1) antibodies on this response. The intradermal injection of recombinant human TNFalpha induced a slowly developing, dose-dependent accumulation of 111In-eosinophils in rat skin that was maximal at the dose of 10(-11) mol/site. Coadministration of TNFalpha with the soluble TNFalpha receptor (p55)-IgG fusion protein (TNFR-IgG) totally inhibited the 111In-eosinophil accumulation induced by the cytokine. The TNFalpha-induced 111In-eosinophil accumulation was not affected after pretreatment of rats with the platelet-activating factor (PAF) receptor antagonist UK-74,505 or the antihuman interleukin-8 monoclonal antibody (MoAb) DM/C7. In contrast, the intravenous administration of an anti-alpha4 integrin MoAb, HP2/1 (3.5 mg/kg), or an anti-VCAM-1 MoAb, 5F10 (2 mg/kg), greatly inhibited the 111In-eosinophil accumulation induced by TNFalpha (the responses detected at 10(-11) mol/site were inhibited by 78% and 50%, respectively). These results show that TNFalpha is an effective inducer of eosinophil accumulation in vivo, with this response being dependent on an interaction between alpha4 integrins and VCAM-1.

The potential role of sialoadhesin as a macrophage recognition molecule in health and disease.

Sialoadhesin is a macrophage-restricted transmembrane glycoprotein of 185 kDa that mediates cell-cell interactions through recognition of Neu5Ac alpha2,3Gal in glycoconjugates. The extracellular region of sialoadhesin is composed of seventeen immunoglobulin-like domains, of which the amino-terminal two are highly-related structurally and functionally to the amino-terminal domains of CD22, myelin associated glycoprotein and CD33. These proteins, collectively known as the sialoadhesin family, are able to mediate sialic acid-dependent binding with distinct specificities for both the type of sialic acid and its linkage to subterminal sugars. In this review we discuss our recent studies on sialoadhesin and suggest how this molecule may contribute to a range of macrophage functions, both under normal conditions as well as during inflammatory reactions.

The species-specific structure of microanatomical compartments in the human spleen: strongly sialoadhesin-positive macrophages occur in the perifollicular zone, but not in the marginal zone.

The microanatomical structure of human and rat splenic white pulp is compared, with special emphasis on the localization of the marginal zone occupied by immunoglobulin M (IgM)+ IgD-/dull B lymphocytes and its specialized macrophages. Our study reveals that in contrast to rats, the marginal zone of humans primarily exists in the vicinity of primary and secondary splenic follicles and that it is almost absent around the periarteriolar T-cell zones. We demonstrate that in humans there is an additional compartment, the perifollicular zone, located between the marginal zone and the red pulp. The perifollicular zone is a dynamic region of variable cellular and phenotypic composition, which can be regarded either as a part of the red pulp or of the follicles. In most cases the perifollicular zone appears as a compartment of the red pulp containing erythrocyte-filled spaces which differ from the typical red pulp sinusoids. Similar to the splenic cords, the perifollicular zone mostly harbours scattered B and T lymphocytes. However, sometimes B lymphocytes clearly predominate in the perifollicular area. In addition, strongly sialoadhesin-positive macrophages form sheaths around capillaries in the perifollicular zone. Such capillary sheaths are not observed in rats. In humans weakly sialoadhesin-positive macrophages are also present in the perifollicular zone and in the red pulp. In some specimens sialoadhesin is, however, strongly expressed by a large number of dispersed perifollicular macrophages. Interestingly, in striking contrast to rats, the human marginal zone does not contain sialoadhesin-positive macrophages and marginal metallophilic macrophages are also absent in humans. Thus, sialoadhesin-positive macrophages and IgM+ IgD- memory B lymphocytes both share the marginal zone as a common compartment in rats, while they occupy different compartments in humans. We show that the human splenic marginal zone does not contain a marginal sinus and assume that in humans the perifollicular region is the compartment where antigen and recirculating lymphocytes enter the organ.

Sialoadhesin (Sn) maps to mouse chromosome 2 and human chromosome 20 and is not linked to the other members of the sialoadhesin family, CD22, MAG, and CD33.

Sialoadhesin is a cell-cell interaction molecule expressed by subpopulations of tissue macrophages. It contains 17 immunoglobulin (Ig)-like domains and is structurally related to CD22, MAG, and CD33. These molecules establish a distinct family of sialic acid-dependent adhesion molecules, the sialoadhesin family. We have mapped the rodent sialoadhesin gene, Sn, to chromosome 2F-H1 by in situ hybridization (ISH) and shown linkage to Il1b and four other markers by backcross linkage analysis. We have also used ISH and a human-mouse somatic cell hybrid panel to localize the human sialoadhesin gene, SN, to the conserved syntenic region on human chromosome 20p13. This demonstrates that the sialoadhesin gene is not linked to the other members of the Sialoadhesin family, CD22, MAG, and CD33, which have been independently mapped to the distal region of mouse chromosome 7 and to human chromosome 19q13.1-3.

Sialoadhesin, myelin-associated glycoprotein and CD22 define a new family of sialic acid-dependent adhesion molecules of the immunoglobulin superfamily.

BACKGROUND: Protein-carbohydrate interactions are believed to be important in many biological processes that involve cell-cell communication. Apart from the selectins, the only well-characterized vertebrate sialic acid-dependent adhesion molecules are CD22 and sialoadhesin; CD22 is a member of the immunoglobulin superfamily that is expressed by B lymphocytes and sialoadhesin is a macrophage receptor. The recent cloning of the gene encoding sialoadhesin has shown that it is also immunoglobulin-like. Both proteins share sequence similarity with the myelin-associated glycoprotein, an adhesion molecule of oligodendrocytes and Schwann cells that has been implicated in the process of myelination, raising the important question of whether myelin-associated glycoprotein is also a sialic acid-binding protein. RESULTS: We have investigated the binding properties of these three receptors when expressed either in monkey COS cells or as chimaeric proteins containing the Fc portion of human immunoglobulin G. We demonstrate that, like sialoadhesin and CD22, myelin-associated glycoprotein mediates cell adhesion by binding to cell-surface glycans that contain sialic acid. We have dissected the specificities of these three adhesins further: whereas sialoadhesin binds equally to the sugar moieties NeuAc alpha 2-->3Gal beta 1-->3(4)GlcNAc or NeuAc alpha 2-->3Gal beta 1-->3GalNAc, myelin-associated glycoprotein recognizes only NeuAc alpha 2-->3Gal beta 1-->3GalNAc and CD22 binds specifically to NeuAc alpha 2-->6Gal beta 1-->4GlcNAc. Furthermore, we show that the recognition of sialylated glycans on the surfaces of particular cell types leads to the selective binding of sialoadhesin to neutrophils, myelin-associated glycoprotein to neurons and CD22 to lymphocytes. CONCLUSIONS: Our findings demonstrate that a subgroup of the immunoglobulin superfamily can mediate diverse biological processes through recognition of specific sialylated glycans on cell surfaces. We propose that this subgroup of proteins be called the sialoadhesin family.

CD69 is expressed by human eosinophils activated in vivo in asthma and in vitro by cytokines.

CD69 is an early activation marker for T cells and cross-linking of CD69 on platelets triggers aggregation and mediator release. Expression of a number of membrane receptors is induced on eosinophils after culture with certain cytokines. Therefore, we investigated whether cytokine-activated eosinophils expressed CD69. Unstimulated, peripheral blood eosinophils did not express CD69, as determined by immunofluorescence and flow cytometry (n = 15). CD69 expression was induced on eosinophils by granulocyte-macrophage colony-stimulating factor (GM-CSF) in a time- and dose-dependent manner. After 1 day in culture, expression was significant at concentrations of 10(-11) M and above. CD69 expression could be detected after stimulation with GM-CSF for only 1 hr, was significant after 2 hr and was sustained over 1-2 days in culture. CD69 expression was also induced by interleukin-3 (IL-3), IL-5 and interferon-gamma (IFN-gamma), but stimulation of eosinophils with platelet-activating factor (PAF) (10(-6) M) for up to 2 hr did not induce CD69 expression. Cycloheximide (10(-6) M) significantly inhibited GM-CSF-induced CD69 expression, suggesting a requirement for protein synthesis. However, unlike up-regulation of CR3 expression, GM-CSF-induced CD69 expression was not inhibited by dexamethasone. CD69 was present on eosinophils from the bronchoalveolar lavage (BAL) fluid of patients with mild asthma (5/5), suggesting that the in vitro findings may have biological relevance in vivo. Therefore, CD69 can be used as a marker of eosinophil activation by cytokines and is a candidate receptor for triggering eosinophil mediator release in the airways in asthma.

Activated memory T helper cells in bronchoalveolar lavage fluid from patients with atopic asthma: relation to asthma symptoms, lung function, and bronchial responsiveness.

BACKGROUND: Bronchial mucosal inflammation and epithelial damage are characteristic features of asthma. Activation of T helper lymphocytes may contribute to this process by mechanisms including the release of cytokines promoting eosinophil infiltration and activation. METHODS: Bronchial washings and bronchoalveolar lavage fluid were obtained from 29 atopic asthmatic patients (19 with current symptoms and 10 symptom free) and 13 normal volunteers. Flow cytometry was used to assess T cell phenotype and activation status in bronchoalveolar lavage fluid and peripheral blood, and differential cell counts were made on bronchial washings and bronchoalveolar lavage fluid. Findings were related to severity of disease as reflected by symptom scores, baseline lung function, and airway responsiveness. RESULTS: CD4 T lymphocytes in bronchoalveolar lavage fluid and blood from asthmatic patients were activated by comparison with controls (CD4 CD25, median 16.8% v 8.7% for bronchoalveolar lavage fluid, and 15.3% v 8.7% in blood). Bronchoalveolar lavage fluid CD4 T cells from both asthmatic patients and controls were of memory phenotype (95.8% and 96.8% CD45RO and 1.7% and 0.4% CD45RA respectively), whereas both CD45RO and CD45RA T cells were present in blood. Patients with asthma and current symptoms showed increased bronchoalveolar T cell activation compared with patients without symptoms (CD4 CD25 18.7% v 12.3%). Within the asthmatic group there was a significant association between CD4 CD25 lymphocytes and asthma symptom scores (rs = 0.75), airway methacholine responsiveness (log PC20, rs = -0.43) and baseline FEV1 (rs = -0.39). A correlation was also found between CD4 CD25 lymphocytes and eosinophils in bronchoalveolar lavage fluid (rs = 0.48). Eosinophils in bronchoalveolar lavage fluid were increased in asthmatic patients compared with controls and the percentage of eosinophils in bronchoalveolar lavage fluid correlated with asthma symptom score. A relation was found between percentage of epithelial cells in bronchoalveolar lavage fluid and FEV1 and methacholine PC20. CONCLUSION: These results support the hypothesis that selective activation of memory CD4 T cells contributes to eosinophil accumulation, bronchial hyperresponsiveness, and symptoms in asthma.

Interleukin-3-induced up-regulation of CR3 expression on human eosinophils is inhibited by dexamethasone.

Eosinophil function is regulated by several cytokines, including interleukin-3 (IL-3), IL-5 and granulocyte-macrophage colony-stimulating factor (GM-CSF). Culture of human eosinophils with IL-3 produced a marked, dose-dependent up-regulation of CR3 expression. This was maximal after 1 day in culture and dependent on protein and RNA synthesis, as demonstrated by inhibition with cycloheximide and actinomycin D, respectively. IL-5 and GM-CSF had a similar effect on eosinophil complement receptor type 3 (CR3) expression, but the maximal response to IL-5 was always less than to IL-3 or GM-CSF. Dexamethasone inhibited the Il-3-induced up-regulation of CR3 expression in a dose-dependent manner, with an IC50 of 5 x 10(-8) M. This study demonstrates the effect of IL-3, IL-5 and GM-CSF on eosinophil CR3 expression and confirms the capacity of eosinophils to modify their phenotype through de novo protein synthesis. This process could be inhibited by physiological concentrations of glucocorticoids, thus providing an additional mechanism for their mode of action in allergic disease.

IFN-gamma induces expression of Fc gamma RIII (CD16) on human eosinophils.

The extent to which eosinophils constitutively express FcRIII (CD16) is controversial. We were unable to detect this receptor on freshly isolated, peripheral blood eosinophils. The capacity of eosinophils to change their Fc gamma R expression in vitro has not been previously demonstrated. Culture with IFN-gamma for 1 to 2 days induced FcRIII expression on eosinophils. This effect was dose-dependent and significant at concentrations of 100 U/ml IFN-gamma and above. Expression of FcRI (CD64) and FcRII (CDw32) was also upregulated. These increases were inhibited by cycloheximide (10(-6) M), suggesting a requirement for protein synthesis, and dexamethasone (10(-6) M). Northern blot analysis demonstrated the presence of FcRIII mRNA in eosinophils cultured with IFN-gamma for 2 days but not in unstimulated eosinophils. By contrast, culture with IL-3 caused an up-regulation of eosinophil FcRII expression but did not induce expression of FcRI or FcRIII. The FcRIII expressed by eosinophils after IFN-gamma stimulation was functionally active, as shown by the triggering of eosinophil membrane depolarization and LTC4 generation by an anti-CD16 mAb. Treatment of IFN-gamma-stimulated eosinophils with phosphatidylinositol-specific phospholipase C reduced FcRIII expression, suggesting that, like neutrophils, eosinophils express the phosphatidylinositol glycan-linked form of this receptor. Therefore, this study demonstrates that IFN-gamma-treated eosinophils express a functionally active, phosphatidylinositol glycan-anchored form of FcRIII.