Pathology Network

From time to time, I'll use of this forum to express my concerns with regard to some common and recurring mistakes, problems, or areas of at least fuzzy logic in pathology publications.  A short, but by no means complete list might include such topics as, over-reaching interpretation of results, confusing association with causation, excessive abbreviations, and the apparently increasing difficulty in producing readable grammatically error-free manuscripts.  I'll start with the flawed concept of determining cell of origin.  This discussion will borrow heavily from an editorial that I published in Modern Pathology about 1995.  In the intervening 15 years many have gotten the point, but some remain to be enlightened.

 

Pick up any journal dealing with human neoplasia, and you will often find comments that this or that cellular phenotypic feature provides insight into the cell of origin, ie. histogenesis, of a neoplasm.  When the putative cell of origin is not a known constituent of the organ or tissue in question, the resultant intellectual machinations evoked to explain this deviant cell line can dwarf all other aspects of the publication. 

 

The early version of the APUD cell theory was a prime offender in this regard.  The concept that normal neuroendocrine cells must arise from neural crest precursors is simply not in keeping with more recent and more detailed observations.  And if normal cells are not so constrained, than neoplastic ones certainly are not.  Neoplastic cells have, unless lost through genetic aberration, a full complement of genetic information and part of their nature is the inappropriate derepression and expression of such information.  The latter expression can be observed directly and tells us a great deal about what the cell is doing, but inferences about where the cell came from are just that...inferences.  Witness in this regard the body of literature regarding aberrent expression of intermediate filaments in a wide variety of neoplasms.
 
Several increasingly contentious examples of "histogenetic-morphogenetic" dissonance will further illustrate this point.  Embryonal rhabdomyosarcomas unequivocally express skeletal muscle differentiation.  Does this mean that they derive from pre-existent skeletal muscle. We would all, I hope, agree that this is highly unlikely, at least in adults where normal skeletal muscle is a fixed, post-mitotic cell population.  In addition, rhabdomyosarcomas, clearly arise in tissue devoid of skeletal muscle.  Indeed, a whole host of soft tissue sarcomas including chondrosarcomas, osteosarcomas and others, arise in sites where similar normal constituents do not exist.

 

Small cell (neuroendocrine) carcinomas seem to occur virtually everywhere and were once considered to derive solely from neuroendocrine precursor cells as per the APUD cell theory.  This led to considerable conceptual distress among authors whenever such tumors were documented in tissues, like the uterine cervix, not demonstrated to contain normal neuroendocrine precursor cells.  The clear-cut demonstration in the last several decades of divergent differentiation in such tumors, with mixtures of squamous, glandular, and neuroendocrine features, often in the same cell, has led to the general agreement that such tumors do not need to derive from a neuroendocrine precursor.  Freed from this conceptual constraint, we can now ask more important and potentially answerable questions such as, "Why do so many highly malignant, rapidly growing, small cell neoplasms express neuroendocrine features?"  Are they in some manner recapitulating the prominent neural and neuroendocrine differentiation seen early in embryonic development?  (ie. does "oncology" recapitulate ontogeny?).  Said differently, are the genes derepressed to produce the rapid growth and cytoplasm-poor features of these cells linked to genes of neuroendocrine differentiation because this pattern has ontogenic significance?

 

Finally, even the medullary carcinoma of the thyroid gland, once believed to be solely the offspring of the parafollicular or C-cell, has come into question, though many current discussions continue to indicate that this is a tumor derived from C-cells.  The recognition of tumors with both follicular and parafollicular features has suggested that medullary carcinomas may not invariably derive from C-cells.  Indeed, the reputed origin of the C-cell, itself, from the neural crest has been brought into serious question.
 
A wag in the pre-genetic era once said that cell of origin, like paternity, was a matter of faith.  With the advent of genetic testing, paternity is now a matter of fact.  Cell of origin, however, remains elusive and of questionable value.  We are observing cells at a single point in time, and the observations we are making directly relate only to the differentiation of those particular cells.  The features of their neoplastically transformed "parent" cannot be observed since this cell was destroyed with the first mitotic division.  We can make only one unequivocally correct statement regarding the cell of origin of any neoplasm, we just need to go back far enough in time.  Every neoplasm definitely derived from its host zygote!