show higher levels of hCG secretion than AF‐type AFC cultures.576 Both AF‐ and F‐type AFCs express human lymphocyte antigen (HLA) class I (HLA‐ABC) but not class II (HLA‐DR) surface antigens.577
Extracellular matrix (ECM) studies and other work63, 578–580 defined a number of qualitative differences between AF‐ and F‐type AFCs. The differences in the types of procollagens produced were such that Bryant et al.581 could use these as markers in fusion studies involving AFCs and postnatally derived skin fibroblasts. Johnston et al.582 provided additional evidence for the distinctiveness of the AF cell type (Figure 3.6). Several polypeptide spots were qualitatively different among F and AF clones (see arrows in Figure 3.6). It furthermore provides convincing evidence for a close ontogenetic relationship between E and AF cells, since their two‐dimensional polypeptide patterns are nearly identical (see Figure 3.6).
Figure 3.6 Selected landscapes from two‐dimensional 35S‐methionine‐labeled polypeptide maps of F‐, AF‐, and E‐type total cell homogenates. All three clone types are derived from a single amniotic fluid specimen to exclude genotypic differences as a source of the apparent protein map differences. Horizontal dimension: isoelectric focusing; vertical dimension: polyacrylamide gradient gel electrophoresis. For technical details, see Johnston et al.582 Arrowheads mark consistent differences of polypeptide spot patterns in the vicinity of the easily identifiable actin cluster (A).
Intermediate filament system
The availability of antibodies to and the electrophoretic characterization of components of the cellular cytoskeleton were extended with great success to cultivated AFC types. For example, the close relationship between AF and E cells received support from immunofluorescence studies using antibodies against epidermal keratins.583, 584 Such immunofluorescent staining of keratin filaments also confirmed the epithelioid nature of most cells in AFC cultures.585 However, AF cells (labeled E1 by Virtanen; see Table 3.7) appeared to express intermediate‐type filamentous structures that reacted with both prekeratin and vimentin antibodies. The conclusions from these early studies must be viewed in the context of the limited specificity (mostly to epidermal keratins) of the antibodies then available. Later, Moll et al.586 provided a comprehensive catalog of well‐characterized prekeratin peptides. This new knowledge was then applied to the identification of AFC clones.
Ochs et al.587 found that both AF and F cells coexpress prekeratin and vimentin filaments, and the cytoplasmic margins of a singular cell type lit up strongly with desmoplakin‐specific antibodies (Figure 3.7). These large, polygonal cells, labeled ED cells, have a distinctive cobblestone pattern, a low growth rate, and resistance to trypsin (see Table 3.7). They were referred to as sheath‐like cells by Hoehn et al.556 Coexpression of cytokeratin and vimentin filaments appears to be promoted by serial culture in many cell types of epithelioid origin. Ochs et al.587 referred to the ED cell as archetype E cell, as it retains close cell‐to‐cell contacts by virtue of an abundant number of desmosomes. All other AFC E cell types, and notably AF and F cells, have lost their desmosomes, together with a number of prokeratin peptides. They display only a remnant pattern of cytoskeletal structures.586
Figure 3.7 Immunofluorescence staining of ED‐type amniotic fluid cells using antibodies against desmoplakin. Bar = 0.05 mm. Note the exclusive reaction with cell boundaries (desmosomes). Source: Ochs et al. 1983587. Reproduced with permission from Elsevier.
F‐type AFCs share many properties with classic fibroblast‐like cells from postnatal skin or foreskin: shape, whorl clone pattern, production of collagenous matrixes, failure to produce hCG, ultrastructure, types of surface glycoproteins and remarkable longevity. Figure 3.8 shows that serially propagated derivative cultures of individual F, E, and AF clones show major differences in their longevities.588
Figure 3.8 Serial propagation and longevity of mass culture progeny of F‐, E‐ and AF‐type amniotic fluid cell clones isolated individually from 20 consecutive amniotic fluid specimens (18 weeks gestational age). The number of primary isolates of each clone type is given in parentheses. Note the relative paucity of F‐type isolates. The progeny of F‐type clones, however, reached the greatest number of cumulative population doublings. In contrast, all E‐type isolates were short‐lived, whereas AF‐type isolates display a wide range of longevities. Source: Based on Hoehn et al.556
The origin of colony‐forming cell types
Sites of origin of colony‐forming AFCs that are not at variance with either cytokeratin findings or anatomic considerations include fetal skin, the bronchopulmonary tract, and the collecting ducts of the kidney.589 The latter site is of particular interest because kidney tissue has been implicated as a source of trisomy 20 cells,590 although trisomy 20 has also been identified in fibroblasts cultured from foreskin.591 Cells staining with an antibody to glial fibrillary acidic protein (GFAP) occur in native fluids, even in the absence of NTDs, but apparently do not form proliferative colonies.592 Enzyme expression593–596 and morphologic resemblance to either fetal urine‐derived cells597 or amnion‐derived cells555 were the early clues to the possible sites of in vivo origin of these cells. hCG, normally produced by the placenta, appeared to be produced by AF‐type but not by F‐type cells in culture.571, 574–576, 576 These studies suggest that the amniotic membranes contribute to the pool of proliferating AFCs.599 Harris600 arrived at a similar conclusion based on her studies of glycoproteins secreted by AFCs.
Subsequent cytoskeleton studies contradicted these earlier findings. Regauer et al.589 found that in situ and cultivated amniotic membrane cells display a much higher cytokeratin structural complexity than any of the AF‐derived cell types, and considered the amnion an unlikely source of clonable cells. They also failed to find concordance between the cytokeratin pattern of urothelial cells and AFCs. Fetal urine cells likely also contribute to the AFC population. Several studies have shown that human fetal and postnatal urine contains cells
