Commentary on Chapters ‘Clinical and Developmental Aspects’ and ‘Stress Responses of the Adrenal Medulla’

The division of the contributions to this volume into Chromogranins, Ion Channels, Secretion Mechanisms, Clinical and Developmental Aspects, and Stress Responses of the Adrenal Medulla is somewhat arbitrary: there is clearly much overlap among these domains of chromaffin cell research. This commentary covers both Clinical/Developmental, and Stress Responses together, mainly to emphasize this overlap, and the translational importance of these two aspects of the chromaffin cell as a stress transducer.

Genetics, development and stress were three topics of particular clinical relevance that were developed in depth through contributions presented at the 15th International Symposium for Chromaffin Cell Biology. The genetics of the chromaffin cell in normal physiology and disease were highlighted by contributions focusing on markers of chromaffin cell origin relevant to cancer including pheochromocytoma, and on catecholaminergic markers for hypertension and other cardiovascular diseases.

Colon, pancreas, prostate and lung outrank the adrenal medulla as organs of medical concern in human cancer. Although pheochromocytoma, or cancer of adrenomedullary origin is rare, it is an important subject of study for several reasons. First, it arises from a limited and distinct set of cell types, and therefore is an attractive model for etiology of proliferative disease. Second, pheochromocytoma can be either metastatic, malignant or benign, and therefore can be used to search for markers to make this clinically crucial diagnosis. Murthy et al. propose carboxypeptidase E (CPE), the enkephalin prohormone processing enzyme first characterized in bovine chromaffin granules in 1982 (Hook et al., Nature 295:341–342, 1982), as such a marker. Meta-analysis of microarray expression data sets in the Gene Expression Omnibus (GEO) allowed Murthy and colleagues to identify CPE as a transcript frequently represented in the transcriptomes of metastatic cancers of both endocrine and epithelial origin. Thus, cervical, colorectal, renal, and bone (Ewing sarcoma), astrocytic, and oligodendroglial tumors or cells express higher levels of CPE mRNA than their non-cancerous cells and tissues of origin: in fact most of the corresponding cells of origin do not have appreciable concentrations of CPE. Lung, pituitary and pheochromocytoma tissue (i.e. neuroendocrine tumors) express higher CPE levels than their tissues of origin. Most significantly, metastatic pheochromocytoma expresses significantly higher CPE mRNA than benign. Should CPE prove a reliable prognostic marker for malignancy in pheochromocytoma and other neuroendocrine cancers, and perhaps even in non-endocrine cancers, its potential for translation to standard clinical practice would be high.

Thouennon and Anouar report on expression of neuropeptides in pheochromocytoma that may be involved in trophic, proliferative and angiogenic paracrine/autocrine actions contributing to tumor growth. A strong correlation between NPY and PACAP expression was documented in 25 pheochromocytomas, and between RDC1, the adrenomedullin receptor, and VEGF, with angiogenic activity. These authors review the evidence for concerted expression of trophic peptides and their receptors in pheochromocytoma leading to autocrine/paracrine regulation of tumor cell growth. An intriguing possibility is that there is ‘cross regulation’ of heterologous peptides and their receptors such that a complex coordinated growth program is initiated and maintained.

O’Connor and colleagues provide two in-depth examinations of genetic effects of catecholamine-related genes on renal hypertension and stress-induced blood pressure regulation. The genetic effects of allelic variation in the tyrosine hydroxylase (TH) gene reviewed by Rao et al. trace the linkages from genetic variation to functional variation, with a face-valid hypothesis for penetrance to altered phenotype resulting in disease (hypertension). The genetic effects of allelic variation in the chromogranin A (CHGA) gene reviewed by Chen et al. are more complex; exhibited at multiple loci in the CHGA gene, some of which are functional (affecting transcription of the CHGA mRNA or processing of the CHGA protein); and complicated in extension to phenotype and disease by the phenomenon of heterosis. TH is the rate-limiting enzyme for catecholamine production. Therefore the identification of a tetranucleotide repeat in the first intron of the gene with two common variants (TCAT6 and TCAT10) that predict autonomic traits in twins provides a straightforward opportunity to connect TH allelic variation to blood pressure and blood pressure response variation through variation in catecholamine production in human individuals. Indeed TCAT9 is associated with higher plasma norepinephrine levels. However, there is not a simple relationship between TCAT number (variation is commonly from six to 11 copies) and hemodynamic responses to stress or resistance to cardiovascular disease. A second polymorphic block exists in the TH promoter region, and affects transcriptional activity, possibly through variable binding of transcription factors (MEF2, SP1, AP2, EGR1, SRY and FOXD1 are candidates). Variants with the greatest effects on TH transcription correspond to individuals with highest catecholamine levels in vivo. At least one allele (C-824T) shows augmented secretory response in cell culture models at nicotinic and PACAP receptors, i.e. to the two co-transmitters at the splanchnicoadrenomedullary synapse. The linkage between TH allelic variation, TH transcription, catecholamine levels and cardiovascular disease is not always simple (i.e. allelic variation blocks corresponding to greater TH transcription, corresponding to higher catecholamine levels, corresponding to hypertension, corresponding to disease). This is not surprising: the “area under the curve” representing integrated sympathoadrenal enhanced catecholamine secretion, and recovery, following stress may be optimized for homeostasis (or for disease-associated allostasis) at several points of regulation. Lack of simple correlation between disease and transcription rate of the disease-linked gene is more pronounced in studies of genetic variants in CHGA linked to hypertension and hypertension-associated kidney disease. Variants in the CHGA promoter linked to glomerular filtration rate (GFR, an index of kidney function and dysfunction) and in the 3’ untranslated region of the CHGA gene linked to end-stage renal disease are indeed ‘functional’, i.e. associated with altered gene transcription of correspondingly mutated reporter genes in cells in culture. Moreover, CHGA promoter variation also predicts plasma endothelin-1 levels, and chromogranin A releases this factor along with angiopoietin-2, and von Willebrand Factor from cultured endothelial cells. Thus, along with its role in the biogenesis of catecholamine-containing secretory vesicles, and various direct hormonal effects of CHGA processed protein fragments (see Commentary of K. Helle on the Chromogranin chapter of this volume) throughout the cardiovascular system, chromogranin A also affects the secretion of key regulatory hormones from the endothelium of the vasculature itself. The genetic studies reviewed by Rao et al. and Chen et al. suggest the complexity underlying stress transduction by the chromaffin cell of the adrenal medulla that leads to homeostasis and contributes to allostatic load throughout the lifespan (see Goldstein). Importantly, creating tighter association between disease outcome and precursor or intermediate phenotypes that can be measured simultaneously and often across time, will provide a statistical framework for broader integration of multiple genes and their allelic variants into the mosaic of human diseases which are progressive, predictable and therefore preventable.

Zeniou-Meyer and colleagues review the cellular and molecular biology of Coffin-Lowry, an X-linked skeletomuscular developmental and mental retardation syndrome which is associated with mutations in a single gene, RPS6KA3, encoding the protein kinase RSK2. An RSK2-KO mouse model with corresponding deficits in long-term spatial memory and learning has been found by this group to show long-term potentiation (LTP) synaptic deficits in the amygdala. However, it was not possible in the complex milieu of the brain to clearly identify whether the role of RKS2 in LTP was pre- or post-synaptic. They therefore turned to the PC12 cell model to determine how RSK2 hypofunction might affect secretion. First, using pharmacological inhibition of RSK kinase function, and RNA interference in PC12 cells, it was shown that RSK2 is indeed required for depolarization-induced secretion in PC12 cells (Zeniou-Meyer et al., Proc. Natl. Acad. Sci. USA 105:8434, 2008). RSK2 function in exocytosis appears to be through the phosphorylation of phospholipase D1 (PLD1). Thus, an important new link in regulation of secretion has been identified, and it is one in which malfunction is clearly associated with disease. The next step is to show that RSK2’s role in corticoamygdalar synaptic communication is mediated by the same or a similar mechanism to that found in PC12 cells. Similarities between secretory mechanisms in highly accessible model systems and at difficult to study synapses of the brain is always welcome news in the context of translational research, from the standpoint of both identification of drug targets and high-throughput identification of drug candidates. This is an important advance for understanding and potentially treating Coffin-Lowry patients. It also adds another mental retardation syndrome to the list of those, including Rett’s syndrome, Down syndrome and neurofibromatosis (Ehninger et al., Neuron 60:950, 2008), now seen essentially as reflecting a potentially correctable synaptic transmission defect, rather than an irreversibly misbuilt brain.

Cell model systems are necessary for progress in translational medicine, whether modeling the behavior of the identical cell within an intact organ, or a gene within a cell, or using an accessible cell such as the bovine chromaffin cell as a model for other post-mitotic neural cells which cannot be obtained in sufficient numbers or purity for biochemical experiments. However, a thorough understanding of the human neuroendocrine cell, and the properties of its surrounding environment in a functioning organ, is also critical. Three contributions in this section address the development and properties of the chromaffin cell within the adrenal gland, and the human chromaffin cell in particular. Perez-Alvarez and colleagues review the embryonic origin, development, morphology and physiology of the human adrenal medulla. They contrast, in particular, the stimulus-secretion coupling process in the human compared to mouse, rat and cow adrenal medulla with respect to type and composition of nicotinic ACh receptors, voltage-dependent calcium channel (VDCC) types, and muscarinic receptors. Alpha-7-subunit-containing nicotinic receptors are now identified in human chromaffin cells. Although it is known that PACAP releases catecholamines from fetal human adrenal chromaffin cells, the presence of PACAP as a co-transmitter with ACh in human adrenal gland is as yet unconfirmed. The 25-year history of human chromaffin cell autologous transplantation to the brain for treatment of Parkinson’s disease is described, and the obstacles to success, including low survival and dopamine production are reviewed. Interestingly, chromaffin cell transplantation to spinal cord for pain may be more promising, with therapeutic effect imputed to enkephalin secretion in situ. It is not clear whether ACh or PACAP, both relatively abundant in spinal cord, is the relevant secretagogue for chromaffin cells at that location. Although adult bovine chromaffin cells are of great value as a neuroendocrine cell model because they are post-mitotic, a very small percentage—like adult neurons of the CNS—are progenitor cells and can be isolated in culture. The therapeutic promise of adult human chromaffin cell progenitors grown as chromospheres in culture is explored by Ehrhart-Bornstein et al. Ehrhart-Bornstein and colleagues point out the advantages of culturing adrenal progenitor cells for autologous transplantation in Parkinson’s disease including: (i) the potential for expansion of the cell population, (ii) the expected lack of degeneration in situ seen with mature chromaffin cells following transplantation to brain and (iii) the ability to encourage a dopaminergic phenotype rather than an adrenergic one during cell expansion in culture. Characterizing the cell population as one with growth potential and phenotypic plasticity is an important first step, and this group reports on the existence of developmentally specific transcription factors, including Sox1 and Mash1, in their bovine chromosphere cultures. Given that bovine chromospheres have increased dopamine (DA) secretion after treatment with ascorbic and retinoic acid, the use of similarly treated human chromosphere cultures in autologous transplantation for treatment of Parkinson’s disease is an exciting prospect.

The contribution of Guerineau and colleagues is an important link between developmental aspects of chromaffin cell function considered above, and the properties of the adrenal medulla as a stress transducer, which the remainder of the contributions to this chapter address directly. The authors review the developmental and stress-induced changes in cell–cell communication in the adrenal medulla. They point out that a developmental switch occurs in the adrenal medulla such that the newborn gland features robust intercellular gap junction communication and relatively weak nicotinic cholinergic synaptic transmission. This is reversed during postnatal maturation of the cholinergic innervation of the gland, so that weak cell–cell junctional coupling and strong nicotinic signaling characterizes the adult adrenal medulla. During chronic stress in the adult animal, however, gap junction-mediated intercellular communication seems to be disinhibited. This would allow amplification of signal transduction primarily coming from the splanchnic nerve, in addition to the increased ACh and PACAP release during increased firing, through synergy with gap junction propagation of cellular excitation and subsequent catecholamine secretion. The involvement of gap-junctional plasticity in the effects of chronic stress is an area of potentially fruitful further study.

Briefly describing David Goldstein’s review of adrenal responses to stress (a keynote talk at the 15th ISCCB) at this point in this commentary somewhat ‘buries the headline’ of a major theme of the meeting: the chromaffin cell as a stress transducer. Goldstein traces the concepts of sympathoadrenal stress developed by Cannon and of HPA axis activation as developed by Selye. He points out that a more modern formulation of these two axes in response to stress is that not only do the sympathoadrenal and hypothalamohypophysialadrenocortical (HPA) axes differ in response to specific stressors, but that the sympathetic and adrenomedullary responses are often quite different. Thus metabolic stress—hypoglycemia—preferentially activates the HPA and adrenal medulla while cold stress (prior to actual lowering of core temperature) activates sympathetic outflow preferentially to HPA and adrenomedullary activation. This updating of earlier (but, as Goldstein points out, still widespread) notions about the ‘unitary’ or generalized aspects of stress response is important. It allows a more detailed accounting of human response to stressors that are predominantly sympathetic, and those that are mainly HPA/adrenomedullary. A systems overview of stress and distress from this more detailed perspective is presented in this review. This perspective also makes incorporation of the concept of allostatic load—the ‘wear and tear’ cost of maintaining homeostasis under conditions in which significant adaptation is required—more conveniently applicable to analysis of how the chromaffin cell might be involved in adaptation, and maladaptation, to prolonged stress. In particular, this systems analysis may help shift focus on stress research from ‘sympathoadrenal’ versus HPA to understanding how the hypothalamus coordinates, perhaps through the parvocellular paraventricular nucleus and CRH release, the activation of the adrenal gland as a whole—both cortex and medulla—as a stress-transducing organ.

Ait-Ali et al. further develop the notion of the adrenal gland as a stress transducer not only at the immediate level of enhanced catecholamine secretion, but also increased production of a broader array of secreted informational molecules, including cytokines, growth factors and neuropeptides, than previously appreciated. Pituitary adenylate cyclase-activating polypeptide (PACAP) is the splanchnicoadrenomedullary transmitter that mediates catecholamine biosynthesis and secretion from the adrenal medulla in response to both metabolic and psychogenic stress (Hamelink et al., Proc. Natl. Acad. Sci. 99:461, 2002; Stroth et al., Neuroscience 165:1025, 2010). Based on these findings, Ait-Ali and colleagues (the author of this commentary is a co-author of this contribution) examined by microarray analysis more than 15,000 transcripts potentially regulated by PACAP in cultured bovine chromaffin cells. Of the 370 or so up-regulated transcripts, more than 10% encoded informational molecules—secreted proteins such as cytokines, growth factors, and neuropeptides that act as first messengers at specific cellular receptors—and enzymes that mediate their intracellular and extracellular activation. These findings imply that the stress response is not limited to the known hormones (catecholamines) and neuropeptides that the chromaffin cell secretes. Rather, it engages a transcriptional program that greatly increases the variety and amount of signaling molecules that the adrenal medulla uses to convey the stress response to other organs.

Wong et al. review the role of HIF1alpha as a transcription factor important in mediating the actions of the inducible transcription factor Egr1 and the ubiquitous transcription factor Sp1 on PNMT gene activation. Based on work primarily in PC12 cells from their own laboratory, this group has determined that HIF1alpha mRNA is up-regulated by hypoxia (5% O₂ for 24 h) in these cells, as are Egr1, Sp1 and PNMT. The authors postulate that HIF1alpha, activated by hypoxia, then activates the downstream transcription factors Egr1 and Sp1 to elevate PNMT gene transcription. Although these experiments were carried out in cells in culture using reporter gene assays, the PC12 cell model allowed facile deployment of siRNA for HIF1alpha to establish the functional order of induction of these factors by hypoxia, i.e. the dependence of Egr-1 and Sp1 protein elevation, and PNMT transcription, on HIF-1alpha (Tai et al. J. Neurochem. 109:513, 2009; Tai et al., Brain Res. doi:10.1016/j.brainres.2010.07.036). These findings are of great interest in further understanding of chromaffin cell stress transduction for several reasons. First, these data reveal HIF-1alpha as an inducible transcription factor in PC12 cells. As the authors point out, HIF-1alpha regulation in other cell systems occurs predominantly at post-translational points of regulation. Second, induction of HIF-1alpha was observed in rat adrenal medulla after immobilization stress or PACAP treatment in vivo. This implies, since PACAP has been shown to be a major regulator of adrenomedullary stress responses in vivo (vide supra) that HIF-1alpha is a stress transducer in chromaffin cells not only for hypoxia, but for other metabolic, and for psychogenic stressors. Multiple pathways including protein kinase A, p38, ERK, PKC, IP3 and Akt (protein kinase B) appear to be involved in this induction, consistent with the pleiotropic nature of PACAP signaling in chromaffin cells. Further analysis of this pathway for PNMT induction in vivo and in bovine chromaffin cells will likely produce further insights into the nature of stress transduction by the adrenal gland, and ways in which it might be modulated in sustained stress responding leading to disease (see Goldstein).

The extent of adrenomedullary plasticity during stress is highlighted further by the report of Tillinger et al., demonstrating that the vesicular monoamine transporter 2 (VMAT2) is induced by immobilization stress in PNMT-positive chromaffin cells in vivo. Vesicular uptake of catecholamines ismediated by two isoforms of VMAT, VMAT1 and VMAT2 (Slc18A1 and Slc18A2). Only VMAT2 is expressed in neurons of the central nervous system, while VMAT1 is expressed in chromaffin cells, enterochromaffin cells and SIF cells of all species examined to date. Interestingly, all cells in both human and rhesus macaque express both VMAT1 and VMAT2 (Erickson et al., Proc. Natl. Acad. Sci. USA 93:5166, 1996; Anlauf et al., J. Histochem. Cytochem. 54:201, 2006), while rodent chromaffin cells express mainly VMAT1, with only low levels of VMAT2. Tillinger et al. show that in the rat, VMAT1 is expressed in all chromaffin cells, while basal VMAT2 is confined to norepinephrine (TH+PNMT−) cells and absent from epinephrine (TH+PNMT+) cells. Upon immobilization stress, VMAT2 mRNA levels are increased in the adrenal medulla, and this occurs via de novo expression in epinephrine-containing cells. This induction, which occurs after a single episode of immobilization stress, provides a mechanism for sustained plasticity of chromaffin cells. A copious literature now suggests that altered vesicular transporter expression in CNS is a physiological regulator of quantal size. It is instructive that this mechanism may also serve to augment epinephrine secretion in the adrenal gland responding to stress. The effects of VMAT2 alteration on quantal size and overall integrated secretion of epinephrine will be an interesting question to approach quantitatively and genetically in the mouse in vivo, if VMAT2 induction after stress can be demonstrated in that species as well.

Douglas et al. reviewed cytokine interactions with chromaffin cells guided by the key observation that “[s]ignaling from the activated immune system is a little explored but potentially important non-neural adrenomedullary regulator.” The effects of interferon alpha and gamma on chromaffin cell tyrosine hydroxylase phosphorylation and apoptosis, respectively, of TNF-alpha through TNFR2 on chromaffin cell gene expression, and the effects of IL-1 on augmentation of catecholamine secretion and biosynthesis are all examples. These effects are mediated through distinct pathways: ERK and STAT1 for IFN-alpha, NF-kB for TNF-alpha, and ERK, nitric oxide, guanylate cyclase and NPY as intermediates for IL-1 signaling. How cytokine signaling can be important in adrenomedullary function and stress transduction is perhaps most clearly indicated by the reviewed literature on the anti-inflammatory factor IL-6: in vivo adrenomedullary catecholamine secretion up-regulates plasma IL-6 concentrations, and adrenalectomy reduces stress-induced IL-6 elevation more than 80%. Thus centrally-mediated IL-6 down-regulation could occur via control of splanchnicoadrenomedullary output, and the adrenal itself may be a systemic source of this important cytokine. The isolated chromaffin cell remains a key model for understanding the mechanistic details of this signaling as a basis for assessing efficacy, and potential complications, of new treatments for major and common threats to human life such as sepsis. Douglas et al. point out also that cytokine signaling to the chromaffin cell provides key insights into their cellular actions on neuroendocrine, in contrast to immune, cells with direct application to elucidating mechanisms of signaling in sickness behavior mediated through cytokines in the central nervous system, and the grave CNS side effects of interferons given as drugs in cancer, and elevated in aging and during stress.

The contributions to this chapter establish that the chromaffin cell remains the best simple model of a synapse in the nervous system in which transduction of a presynaptic signal to a functional, physiologically measurable output can be visualized in detail, and in which translation to improved clinical practice can also be clearly envisaged.