| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Letters to the Editor |
The University of Birmingham Birmingham B15 2TG United Kingdom
We read the paper by Kauma et al. (1) with interest. Their technique for separating and culturing trophoblast from villous core tissue showed that only the villous core tissue produces hepatocyte growth factor (HGF). This confirms evidence from previous in situ hybridization studies by ourselves and others, where HGF mRNA was localized to the villous core (2, 3, 4). They attribute the HGF production to villous core fibroblasts but are unable to explain why cultures of whole tissue produce 24 times more HGF than isolated villous core tissue: co-culture of trophoblast cells and isolated villous core fibroblasts failed to increase HGF production. One possible explanation for this lies in our in situ hybridization studies, which clearly show that by far the strongest signal for HGF came from the perivascular smooth muscle of the stem villi. If their technique for isolating villous core tissue resulted in damage or loss of perivascular tissue, then it would not be surprising to see a dramatic reduction in HGF production. Vascular smooth muscle has previously been shown to be a potent source of HGF (5). Furthermore, in Fig. 3 of the Kauma paper, it appears that the whole tissue culture supernatant contained around 240 ng HGF per µg total protein. This implies that nearly 25% of total secreted protein was HGF, which is implausibly high and raises questions about the measurements reported.
The failure by Kauma et al. to immunolocalize HGF to the villous trophoblast is at odds with the previously published work of ourselves and others, showing co-localization of HGF and c-met protein to the vasculo-syncytial membrane in the third trimester (3, 6). One study in human placenta from the first trimester also failed to localize HGF to the trophoblast, but did localize it to the basal membrane area, where the cytotrophoblasts come into contact with the villous core (4). A fourth study failed to immunolocalize HGF to trophoblast throughout gestation (2). The reasons for this discrepancy may well lie in the specificity of the antibodies involved as, while the latter study (Clark et al., ref. 2) and Kauma et al. used the same monoclonal antibody, the other three studies all used different polyclonal antibodies. It is possible that the polyclonal antibodies may bind to forms of HGF not detected by the monoclonal antibody, which is likely to recognize a single epitope. Alternative forms of HGF likely to exist in the trophoblast layer include the inactive monomer, the active dimer, and the receptor bound or internalized HGF. A further difference identified between these studies lies in the tissue fixation used: both Wolf et al. (6) and ourselves (3) used Formalin-fixed wax embedded sections, whilst the other three studies used frozen sections, post-fixed in acetone. This may also affect antigen recognition, especially by monoclonal antibodies. It would be interesting to know if the authors were able to perform successful HGF immunolocalization using the Formalin-fixed, wax embedded sections that they used for the c-met immunohistochemistry.
Footnotes
Received October 1, 1997. Address correspondence to: D. A. Somerset, Academic Department of Obstetrics and Gynecology, Birmingham Women’s Hospital, Edgbaston, Birmingham, United Kingdom B15 2TG.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
