Pituitary Adenoma Differentiation

A further characteristic that distinguishes pituitary adenomas from other tumors is that pituitary adenomas are well differentiated and often remain responsive to normal physiologic stimuli. It is unusual for the secretory identity

Fig. 3. Dark-field photomicrograph of tissue fragments at the edge of a 12-^m-thick frozen section of a somatotroph adenoma hybridized to an oligonucleotide probe complementary to TRH. A subset of individual cells and small clusters of cells expressing TRH at high level are clearly seen.

of the presumed cells of origin to fluctuate and the efficacy of somatostatin and dopamine analogues in somatotroph and lactotroph adenomas and the effects of changes in levels of circulating glucocorticoids in corticotroph adenomas attest to their continued expression of hormone receptors. Progressive deregulation of pituitary adenoma functional activity during propagation and how colonic neoplasms acquire a sequential series of genetic defects over time and become increasingly malignant is unusual. Indeed, although thyrotroph adenomas, for example, may become more resistant to somatostatin analogs over time, the hormone secretory potential of silent corticotroph adenomas may resume years after the original diagnosis and secreting corticotroph adenomas may alternate between quantitatively normal and abnormal hormone secretion (9-11), findings that are difficult to equate with progressive dedifferentiation.

Despite the belief that they develop from the deregulated division of single cells harboring either point mutations or some degree of chromosomal disruption, pituitary adenomas frequently contain well-defined subsets of cells, with distinct transcriptional activity, expressing not only pituitary but also hypotha-lamic hormones (12-14) (Fig. 3). It is almost as if in these cases differentiation has not just been maintained but new neuroendocrine-differentiated function has been unmasked in subsets of cells. Even when hormone secretion is not continued, the histology of endocrinologically inactive adenomas at light and electron microscopic levels clearly identifies the subset of secretory cells of origin in most cases. In summary, pituitary adenomas are usually well differentiated at diagnosis and retain their differentiation in the long-term.

IS PITUITARY PATHOGENESIS INTRINSIC OR DRIVEN BY EXTERNAL GROWTH SIGNALS?

As mentioned, whether pituitary adenomas arise from intracellular defects or from excessive trophic influences from outside remains central to discussions about pathogenesis. The observation that has been pivotal in guiding opinion resolutely towards intracellular defects—to the extent that pituitary behavior has been all but ignored—is pituitary adenoma clonality. This issue is so important conceptually that it warrants explanation and examination even in such a brief summary as this. The clonality of a cellular expansion is a secure archaeologic tool capable of distinguishing an irreversible and potentially inexorably progressive process induced by an intracellular insult or insults from a relatively excessive but possibly reversible or self-limiting trophic response to stromal or microenvironmental signals (15,16).

The basis of clonality is that during early female embryogenesis, each cell randomly but permanently inactivates genes on either the maternally or the paternally derived X-chromosome by cytosine methylation of promoter regions. Once methylated, the pattern of functional inactivation is stably inherited by the progeny of each cell (17). Thus, all adult female tissues consist of a mosaic of cells, in each of which genes on either the maternal or paternal X-chromosome have been inactivated.

Clonality analysis depends first on being able to obtain the tissue of interest (in this case a pituitary adenoma) from a female patient who happens by chance to be heterozygotic for X-chromosome-linked markers, second on being able to distinguish between different X-chromosome alleles in such a patient (straightforward, given the known X-chromosome-linked markers), and finally, on being able to establish, on the basis of the methylation pattern of the alleles, which of the two X-chromosomes is active. If migration and dispersal of cells between assignment of clonality in utero and postnatal life leads to the formation of a mosaic at a single-cell level and mature tissues do not continue to divide, then, unless motile cells of similar identity and clonality spontaneously cosegregate in differentiated tissue or a polyclonal field spontaneously undergoes differential apoptosis leaving a single clone, a monoclonal cellular expansion in a polyclonal field must represent the progeny of a single cell. Conversely, a tumor that develops through inappropriate exogenous stimulation is likely to be derived from more than one susceptible cell and the resulting mass will be polyclonal.

Thus, the finding of monoclonality in pituitary adenomas is broadly taken to be the most compelling evidence for the neoplastic origin of these lesions, and in published series, approx 90% of the 60 or so pituitary tumors that have been analyzed for clonality have been reported to be monoclonal. The summated findings of studies from five different groups (18-22) show that all of the eight endocrinologically inactive adenomas, two gonadotroph adenomas, three somatotroph adenomas, and one mammosomatotroph adenoma analyzed were monoclonal. Two of the six lactotroph adenomas examined and the single plurihormonal adenoma were apparently polyclonal but were contaminated with normal pituitary tissue, and of the 22 corticotrophs examined, 17 were monoclonal, including five of the six tumors recovered from patients with Nelson's syndrome. On the basis of these studies, which were conducted 10 yrs ago, the search for the pathogenesis of pituitary adenomas has focused almost entirely on primary intracellular pituitary defects.

There are several technical and conceptual problems with clonality studies, however, that may have made the decision to abandon studies of the effects of pituitary architecture, hypothalamic stimulation, and feedback effects on pituitary adenoma pathogenesis and propagation premature.

The strategy used to define clonality relies on qualitative assessments of the consistency of deoxyribonucleic acid (DNA) recovery and probe hybridization to membrane-bound DNA that has been subjected to sequential restriction enzyme digestions. Absence of monoclonality could be dismissed as tissue contamination, and arbitrary cut offs, such as a reduction of more than 80% in the optical density of one hybridization band and less than 40% in the other, or decreases of between 70% and 30%, have to be defined as indicative of clonal skewing or polyclonality respectively (23). Publication bias toward a positive result is thus compounded by considerable technical bias in favor of monoclonality.

Despite their presumed monoclonal origin, many pituitary tumors transcribe (14) and translate (24) multiple pituitary hormones in subsets of cells. The explanations adduced for this observation, if monoclonality is taken at face value, include sample contamination, monoclonal expansion of pleurihormonal precursors, a cell cycle phase effect, or further mutations of the original malignant clone. These interpretations are not supported by the pattern of hormones represented, which include hypothalamic hormones (12-14,25,26), by the lack of evidence of new genomic deletions or mutations appearing with time in the majority of cases (27,28), or by the slow and stable intrinsic rate of cellular turnover, which is too low to allow new subclones to emerge (29).

Aside from these inconsistencies, a further critical possibility may confound the simplistic interpretation of clonality studies, which is that dispersal of cells after trophic assignment during development may not be complete (15,16). There are precedents for exactly this macroscopic monoclonality in the absence of neoplasia in other tissues. Smooth muscle cell proliferation in human myometrium (30) normal aortic smooth muscle and atherosclerotic plaques (31,32)

Neoplasia Cartoon

Fig. 4. Cartoon of the appearance of a non-neoplastic monoclonal expansion as an artefact of preexisting clonal topography. The left hand panel shows five different clones of a cellular subtype (depicted as differently shaded spheres), with distinct but overlapping clonal territories distributed within the pituitary (which is represented here as a 'wireframe' ovoid). The remaining pituitary cell types, which under normal circumstances fill the spaces between them (not shown), ensure that despite the nonrandom clonal topography of the particular cellular subtype shown, a random tissue sample will be truly polyclonal. By displacing intervening cells, expansion of a single clone in response to a physiologic stimulus (right hand panel) might lead to the appearance of skewing towards monoclonality in biopsy analysis, yet be derived from a physiologic response of a preexisting clone of normal cells rather than a single mutant.

Fig. 4. Cartoon of the appearance of a non-neoplastic monoclonal expansion as an artefact of preexisting clonal topography. The left hand panel shows five different clones of a cellular subtype (depicted as differently shaded spheres), with distinct but overlapping clonal territories distributed within the pituitary (which is represented here as a 'wireframe' ovoid). The remaining pituitary cell types, which under normal circumstances fill the spaces between them (not shown), ensure that despite the nonrandom clonal topography of the particular cellular subtype shown, a random tissue sample will be truly polyclonal. By displacing intervening cells, expansion of a single clone in response to a physiologic stimulus (right hand panel) might lead to the appearance of skewing towards monoclonality in biopsy analysis, yet be derived from a physiologic response of a preexisting clone of normal cells rather than a single mutant.

all exhibit macroscopic monoclonality. The aorta, for example, consists of a patchwork of contiguous 4-mm patches skewed to the same allele (32). The same is true of regenerative nodules in cirrhotic liver, which are monoclonal but entirely benign (33), and in bladder urothelium, where it is estimated that200-300 founder cells each give rise to approx 2 million cells arranged in macroscopic monoclonal patches approx 120 mm2 (34). Thus, in several tissues in which relatively pure cellular populations allow clonality to be ascertained, and perhaps in the majority of human tissues, macroscopic monoclonality is the norm.

These observations are extremely pertinent because an unusually strong or prolonged stimulus or a sequence of unusually timed stimuli might induce an excessive but nevertheless physiologic trophic response in one clone of pituitary cells that would lead to its progeny increasingly outnumbering and displacing other interspersed cell subtypes, skewing the clonality pattern in that territory. Depending on the threshold dictated by the techniques used to analyze clonality and the timing and location of pituitary biopsy, it is possible that the responsive cell population would become identifiable in vitro as a monoclonal expansion (Fig. 4).

Unlike a neoplastic monoclonal expansion derived from a single abnormal cell, expansion of a clonal patch might represent an entirely physiologic trophic response that would be expected to be well differentiated and physiologically responsive and demonstrate fluctuations in activity or spontaneous resolution, depending on the nature and duration of the trophic stimulus. This scenario would predict the twin paradoxes of a relatively high long-term failure rate for tumor excision and that partial removal of histologically normal pituitary might produce a functional remission—events that are known to occur in clinical practice (8,35-38). The implication is that monoclonality does not necessarily mean neoplasia in the normal sense of the word.

One further extraordinary recent observation has called simple monoclonality into question. In more than half of recurrent pituitary tumors in which it was possible to repeat examination for loss of heterozygosity after further surgery, the loci originally implicated were heterozygous once again; in other words, the missing chromosomal region had been restored (28). The explanation that the recurrence represented a second entirely independent monoclonal tumor derived from another abnormal clone after complete resection of the first is not consistant with clinical experience because a pituitary adenoma that has been almost entirely resected does not normally recur significantly. Restoration of the original allelic deletion by further mutation is clearly vanishingly improbable. A more likely scenario is that in the cases where the integrity of the genome has been restored in tumor recurrence, the original adenomas were initiated as oligoclonal rather than monoclonal expansions and that clonal interference (i.e., selection pressure) facing the newly established expansion resulted in several of the original clones being suppressed and diluted below detectable levels (39-42). Debulking of the original trophically disinhibited clone allowed another clone to become trophi-cally dominant. A more modest trophic response in a second clone with distinct transcriptional and translational activity, such as cotranscription of pro-opiomelanocortin and growth hormone-releasing hormone or corticotrophin-releasing hormone, could account for the appearance of cellular subpopulations yet be insufficient to mask monoclonality (16).

Conceptual arguments about benign monoclonality aside, the pathogenesis of pituitary adenomas is still widely perceived to be related to relatively benign neoplasia rather than to a qualitatively normal but quantitatively excessive response to physiologic trophic stimuli. Proto-oncogene activation is therefore still a critical prerequisite for pituitary tumor formation.

Was this article helpful?

0 0

Post a comment