Accumulation of hypophysiotropic hormone mRNAs in somatotroph pituitary tumors: Biological, clinical and prognostic implications

The behavior of somatotroph pituitary tumors greatly varies, however, the events/mechanisms underlying this variability remain poorly understood. These tumors arise from transformation of somatotrophs, the growth hormone (GH) producing pituitary cells. The secretory and proliferative capabilities of the normal somatotroph are governed by two hypothalamic (hypophysiotropic) hormones: growth hormone releasing hormone (GHRH) and somatostatin (SRIF). We have shown that the neoplastic somatotroph can, in a nonrandom fashion, assume the capacity to transcribe and translate genes encoding GHRH and SRIF, the mRNA levels of which are associated with prognostically relevant differences in tumor behavior. In 100 adenomas studied by in situ hybridization, the distribution of GHRH and SRIF mRNA transcripts was quantified, and their clinicopathologic correlates determined; protein translation was confirmed by Western blotting. GHRH transcripts were found to preferentially accumulate within aggressive tumors. Specifically, GHRH mRNA signal intensity was: (i) linearly correlated with Ki-67 tumor growth fractions; (ii) linearly correlated with blood GH levels; (iii) higher among invasive tumors; and (iv) highest in those tumors in which postoperative remission was not achieved. Alternatively, SRIF transcripts were found to preferentially accumulate among clinically favorable variants; the SRIF mRNA signal was: (i) inversely correlated with Ki-67 tumor growth fractions; (ii) inversely correlated with preoperative GH levels; (iii) higher in noninvasive tumors; and (iv) highest in tumors amenable to surgical remission. From the GHRH and SRIF mRNA signal intensities, several significant logistic models of postoperative outcome were fitted and validated in a second population of 30 somatotroph adenomas. The pattern of GHRH and SRIF transcript accumulation permitted tumors to be meaningfully grouped into aggressive and favorable variants; aggressive variants expressed high and low levels of GHRH and SRIF transcripts, respectively; favorable variants expressed high and low levels of SRIF and GHRH transcripts, respectively. That the latter clinically favorable state could be approximated pharmacologically with presurgical therapy an SRIF analog. This agent was associated with an antiproliferative effect and reductions in GHRH mRNA and protein expression. The nonrandom and prognostically informative patterns of GHRH and SRIF transcript accumulation supports an autocrine/paracrine regulatory role for these hypophysiotropic hormones in this tumor system.