Two novel human β-chemokines, Ckβ-8 or myeloid progenitor inhibitory factor 1 (MPIF-1), and Ckβ-6 or MPIF-2, were discovered as part of a large scale cDNA sequencing effort. The MPIF-1 and MPIF-2 cDNAs were isolated from aortic endothelium and activated monocyte libraries, respectively. Both of the cDNAs were cloned into a baculovirus vector and expressed in insect cells. The mature recombinant MPIF-1 protein consists of 99 amino acids and is most homologous to macrophage inflammatory protein (MIP)-1α, showing 51% identity. It displays chemotactic activity on resting T lymphocytes and monocytes, a minimal but significant activity on neutrophils, and is negative on activated T lymphocytes. MPIF-1 is also a potent suppressor of bone marrow low proliferative potential colony-forming cells, a committed progenitor that gives rise to granulocyte and monocyte lineages. The mature recombinant MPIF-2 has 93 amino acid residues and shows 39 and 42% identity with monocyte chemoattractant protein (MCP)-3 and MIP-1α, respectively. It displays chemotactic activity on resting T lymphocytes, a minimal activity on neutrophils, and is negative on monocytes and activated T lymphocytes. On eosinophils, MPIF-2 produces a transient rise of cytosolic Ca2+ and uses the receptor for eotaxin and MCP-4. In hematopoietic assays, MPIF-2 strongly suppressed the colony formation by the high proliferative potential colony-forming cell (HPP-CFC), which represents a multipotential hematopoietic progenitor.

A novel LDL-associated phospholipase A2 (LDL-PLA2) has been purified to homogeneity from human LDL obtained from plasma apheresis. This enzyme has activity toward both oxidized phosphatidylcholine and platelet activating factor (PAF). A simple purification procedure involving detergent solubilization and affinity and ion exchange chromatography has been devised. Vmax and Km for the purified enzyme are 170 micromol.min-1.mg-1 and 12 micromol/L, respectively. Extensive peptide sequence from LDL-PLA2 facilitated identification of an expressed sequence tag partial cDNA. This has led to cloning and expression of active protein in baculovirus. A lipase motif is also evident from sequence information, indicating that the enzyme is serine dependent. Inhibition by diethyl p-nitrophenyl phosphate and 3,4-dichloroisocoumarin and insensitivity to EDTA, Ca2+, and sulfhydryl reagents confirm that the enzyme is indeed a serine-dependent hydrolase. The protein is extensively glycosylated, and the glycosylation site has been identified. Antibodies to this LDL-PLA2 have been raised and used to show that this enzyme is responsible for >95% of the phospholipase activity associated with LDL. Inhibition of LDL-PLA2 before oxidation of LDL reduces both lysophosphatidylcholine content and monocyte chemoattractant ability of the resulting oxidized LDL. Lysophosphatidylcholine production and monocyte chemoattractant ability can be restored by addition of physiological quantities of pure LDL-PLA2.

With the powerful deep network architectures, such as generative adversarial networks and variational autoencoders, large amounts of photorealistic images can be generated. The generated images, already fooling human eyes successfully, are not initially targeted for deceiving image authentication systems. However, research communities as well as public media show great concerns on whether these images would lead to serious security issues. In this paper, we address the problem of detecting deep network generated (DNG) images by analyzing the disparities in color components between real scene images and DNG images. Existing deep networks generate images in RGB color space and have no explicit constrains on color correlations; therefore, DNG images have more obvious differences from real images in other color spaces, such as HSV and YCbCr, especially in the chrominance components. Besides, the DNG images are different from the real ones when considering red, green, and blue components together. Based on these observations, we propose a feature set to capture color image statistics for detecting the DNG images. Moreover, three different detection scenarios in practice are considered and the corresponding detection strategies are designed. Extensive experiments have been conducted on face image datasets to evaluate the effectiveness of the proposed method. The experimental results show that the proposed method is able to distinguish the DNG images from real ones with high accuracies. Index TermsImage generative model, generative adversarial networks, fake image identification, image statistics.