Shin-Ichi Nishikawa

PDGF Mediates a Neuron–Astrocyte Interaction in the Developing Retina

Astrocytes invade the developing retina from the optic nerve head, over the axons of retinal ganglion cells (RGCs). RGCs express the platelet-derived growth factor A-chain (PDGF-A) and retinal astrocytes the PDGF alpha-receptor (PDGFR alpha), suggesting that PDGF mediates a paracrine interaction between these cells. To test this, we inhibited PDGF signaling in the eye with a neutralizing anti-PDGFR alpha antibody or a soluble extracellular fragment of PDGFR alpha. These treatments inhibited development of the astrocyte network. We also generated transgenic mice that overexpress PDGF-A in RGCs. This resulted in hyperproliferation of astrocytes, which in turn induced excessive vasculogenesis. Thus, PDGF appears to be a link in the chain of cell-cell interactions responsible for matching numbers of neurons, astrocytes, and blood vessels during retinal development.

Interspecies molecular chimeras of kit help define the binding site of the stem cell factor.

The extracellular portion of the kit-encoded receptor for the stem cell factor (SCF) comprises five immunoglobulin (Ig)-like domains. To localize the ligand recognition site, we exploited the lack of binding of human SCF to the murine receptor by using human-mouse hybrids of Kit and species-specific monoclonal antibodies (MAbs) that inhibit ligand binding. Replacement of the three N-terminal Ig-like domains of the murine Kit with the corresponding portion of the human receptor conferred upon the chimeric receptor high-affinity binding of the human ligand as well as of human-specific ligand-inhibitory MAbs. By constructing five chimeric murine Kit proteins which individually contain each of these three human Ig-like units or pairs of them, we found that the second human domain confers upon the mouse Kit high-affinity binding of the human ligand and also binding of species-specific SCF-competitive MAbs. Nevertheless, the flanking Ig-like domains also affect high-affinity recognition of SCF. Moreover, it appears that the determinants that define ligand specificity of the murine and the human receptors do not structurally coincide. This observation allowed us to identify a chimeric receptor that displayed a dual specificity; namely, it bound with high affinity either the human or the murine SCF molecules and reacted with mouse- as well as human-specific ligand-inhibitory MAbs. Conversely, another chimera, which included all of the five Ig-like domains, bound neither ligand. In conclusion, interdomain packing involving the second Ig-like domain of human Kit and noncontiguous structural motifs of the receptor are involved in SCF recognition.

Neural and skin cell-specific expression pattern conferred by steel factor regulatory sequence in transgenic mice

We have produced transgenic mice expressing a lacZ reporter gene under the control of a fragment of a Steel factor (SLF). The function of this gene is essential for the development of hematopoietic cells, germ cells, melanocytes and pacemaker cells of the intestine. The expression of the transgene, containing 2 kb DNA 5′ regulatory sequence, was restricted to neural and skin tissues in appropriate spatial and temporal pattern compared with endogenous SLF mRNA expression. This indicates that the regulatory elements necessary for the neural and skin specific expression are present in this 2 kb DNA sequence, although strong position‐dependence of transgene expression was observed. As we could not detect transgene expression in hematopoietic tissues and germ cells after extending 10 kb upstream, elements important for these organs must reside in other regions. Our results indicate that neural crest derived enteric ganglion cells provide SLF to the neighboring pace maker cells expressing c‐kit, the receptor for SLF. Cells expressing the transgene in the intestine are ganglion cells derived from neural crest since homozygosity for the lethal spotting (ls) mutation results in loss of such ganglion cells in transgenic mice. We have also shown that the dermal papillae of the hair follicle expresses the transgene, suggesting its roles to support the c‐kit dependent growth and development of melanocytes in the hair follicle.

Constitutive expression of interleukin-7- mRNA and production of IL-7 by a cloned murine thymic stromal cell line

Mouse interleukin‐7 (IL‐7) cDNA was cloned from mouse thymic stromal cell clone MRL104.8a using a polymerase chain reaction (PCR) technique and expressed in COS‐7 cells. The resulting recombinant interleukin‐7 (rlL‐7) supported the proliferation of mouse antigen‐specific helper T cell (Th) clone 9–16 in the absence of IL‐2 and antigen as well as mouse pre‐B cell line DW34. It was also found that high levels of the mRNA for IL‐7 were constitutively expressed in the MRL104.8a cells, and a potent amount of IL‐7 was produced in its culture supernatant. These results provide the evidence for constitutive expression of IL‐7 mRNA and for production of IL‐7 by thymic stromal cells that have a critical role in intrathymic T cell development. The results are discussed in the context of the functional and molecular relationship between IL‐7 and the previously described cytokines produced by thymic stromal cells.

Osteoclast precursors in bone marrow and peritoneal cavity

Osteoclasts differentiate from cells that share some phenotypes with mature macrophages and monocytes, but early precursors for osteoclasts still remain obscure. To characterize osteoclast precursors, using monoclonal anti‐c‐Fms and anti‐c‐Kit antibodies, bone marrow cells were separated and the frequency of clonogenic progenitors were measured. Osteoclast precursors in the bone marrow mainly expressed c‐Kit and diminished in frequency when they expressed c‐Fms. In contrast to bone marrow, the precursors in the peritoneal cavity were enriched with a population of c‐Fms+. Injection of these antibodies into mice demonstrated that peritoneal osteoclast precursors were sensitive to anti‐c‐Fms but not to anti‐c‐Kit antibodies, whereas those in bone marrow only declined in the presence of both antibodies. Meanwhile, c‐Fms as opposed to c‐Kit played an essential role in the generation of osteoclasts in cultures. We also compared osteoclast precursors with colony forming cells (CFU‐M) by a macrophage colony stimulating factor. CFU‐M in bone marrow decreased when anti‐c‐Kit antibody was administered and no CFU‐M was detected in peritoneum. In this study, we show differences between proliferative potential osteoclast precursors maintained in bone marrow and peritoneum and between CFU‐M and osteoclast precursors. J. Cell. Physiol. 170:241–247, 1997.

Enteric neurons express Steel factor-lacZ transgene in the murine gastrointestinal tract

The cells that express Steel factor (SLF) in the gastrointestinal (GI) tract were studied using SLF-lacZ transgenic mice. Expression, detected by beta-galactosidase histochemistry, was evident in cells between the circular and longitudinal muscle layers in the GI tract. Double staining with antibodies specific for the neural markers, PGP 9.5, MAP2 and c-Ret, showed that SLF-lacZ positive cells were enteric neurons. Enteroglia did not express SLF-lacZ. The distribution of expressing cells was complimentary to the expression of c-Kit in myenteric interstitial cells.

Establishment of IL-5-Dependent Early B Cell Lines by Long-Term Bone Marrow Cultures

We established two different IL-5-dependent Ly1+ early B cell lines in long-term bone marrow culture system. One of them (J-87) is stromal cell (ST2) dependent and the other (T-88) is ST2 independent. Both J-87 and T-88 are B220+, Ly1+, sIgM-, Ia-, Thy1-, and IL-2R+, and respond to IL-3 and IL-5 in the presence of ST2. The T-88 can proliferate only in response to IL-5 in the absence of ST2. Southern blot analysis using JH probe revealed that configuration of IgH gene of both cell lines shows rearranged pattern. Binding assay for radiolabeled IL-5 to T-88 demonstrated that T-88 has two classes of IL-5 binding sites (low and high affinity) on the membrane. These data strongly suggest that there are IL-5-sensitive stages at both stromal cell-dependent and stromal cell-independent phases in early B cell development.

The Role of c-kit Proto-oncogene during Melanocyte Development in Mouse. In vivo Approach by the In utero Microinjection of Anti-c-kit Antibody

In order to investigate the role of the c‐kit oncogene in the melanoblast development, a rat monoclonal antibody (ACK2) against the mouse c‐kit protein was used to localize cells expressing c‐kit during fetal development. ACK2 was also injected directly into the amniotic cavity of mouse fetuses at successive developmental stages. After birth, the offspring were examined to determine the resulting coat color patterns. c‐kit positive melanoblasts first appeared in dermis of fetuses at 11.5 days postcoitum (dpc). Subsequently, these cells increased in number and migrated dorsolaterally to the ventral region, and by 12.5 dpc some of them began to invade the epidermis. Treatment of fetuses by ACK2 microinjection appeared to affect the pigmentation in the coat, inducing a variety of spotting patterns in offspring, and the location of the spots was closely correlated with gestational stage. ACK2 injection of early fetuses produced major changes in coat color even though few c‐kit positive cells were detectable in the dermal mesenchyme at the time of injection. Large spots were also induced when mid‐stage fetuses with a only few c‐kit positive cells in the dorsal region were injected. By contrast, except for spot formation in the center of ventral region, ACK2 injection did not appear to affect melanogenesis in late stage fetuses that had many c‐kit positive cells.

Characterization and isolation of melanocyte progenitors from mouse embryos

Whole mount immunohistochemistry and flow cytometry have been used to determine the morphological and molecular features that distinguish melanoblasts from surrounding cells. Whole mount immunohistochemistry of mouse embryos using anti‐c‐Kit monoclonal antibody revealed two distinct types of c‐Kit+ cells; one dendritic and the other round in shape. The distribution of c‐Kit+ dendritic cells in 12.5 days postcoitem embryos correlated well with that of tyrosinase‐related protein‐2 expression, while the distribution of c‐Kit+ round cells overlaps that of CD45+ cells. This observation suggests that melanoblasts are distinguishable from other c‐Kit+ cells by their dendritic shape. Mice homozygous for the steel‐Dickie mutation (Sld/Sld) were analyzed to further distinguish melanoblasts from hematopoietic progenitor cells. Sld/Sld mice are unpigmented but contain hematopoietic cells, although reduced in number. Although no c‐Kit+ dendritic cells were detectable in the Sld/Sld embryos, a significant number of c‐Kit+ round cells were present in the same embryos. To further analyze characteristic features of melanoblasts, c‐Kit+CD45− and c‐Kit+CD45+ cells were isolated from dissociated embryonic skin by fluorescent activated cell sorter and the expression of TRP2 melanogenic enzyme was analyzed. Consistent with histological analysis, most c‐Kit+CD45− cells were TRP2+.c‐Kit+CD45+ cells failed to express TRP2. These results show that most of the melanoblasts are c‐Kit+TRP2+CD45− dendritic cells and can be discriminated from other cells by flow cytometry or by their morphology.

Effects of macrophage colony-stimulating factor (M-CSF) on the development, differentiation, and maturation of marginal metallophilic macrophages and marginal zone macrophages in the spleen of osteopetrosis (op) mutant mice lacking functional M-CSF activity.

Immunohistochemical techniques using an anti‐mouse panmacrophage monoclonal antibody and anti‐mouse monoclonal antibodies specific for marginal metallophilic macrophages or marginal zone macrophages were used to detect red pulp macrophages, marginal metallophilic macrophages, and marginal zone macrophages in the spleen of op/op mice. In the mutant mice, the red pulp macrophages were reduced to about 60% of those in the normal littermates and the marginal metallophilic macrophages and marginal zone macrophages were absent. After administration of recombinant human macrophage colony‐stimulating factor (rhM‐CSF), numbers of red pulp macrophages increased rapidly, reaching levels found in normal littermates 1 week later. In contrast, the marginal metallophilic macrophages as well as the marginal zone macrophages appeared slowly after rhM‐CSF administration and their numbers were less than half of the baseline level of normal littermates even at 12 weeks of administration. The distribution of marginal metallophilic macrophages and marginal zone macrophages appearing after M‐CSF administration was irregular in the spleen of the op/op mice. These splenic macrophage subpopulations differed in their responses to rhM‐CSF, suggesting that distinct mechanisms may be involved in their development and differentiation. The splenic red pulp macrophages present in unmanipulated op/op mice are an M‐CSF–independent macrophage population. Although the marginal metallophilic macrophages and marginal zone macrophages are thought to be M‐CSF‐dependent, their development and differentiation appear to be influenced by locally produced M‐CSF or other cytokines. J. Leukoc. Biol. 55: 581‐588; 1994.