The mitochondrial component of intracrine action

Abstract

In recent years the actions of intracellular-acting, extracellular signaling proteins/peptides (intracrines) have become increasingly defined. General principles of intracrine action have been proposed. Mitochondria represent one locus of intracrine action, and thus far, angiotensin II, transforming growth factor-β, growth hormone, atrial natriuretic peptide, Wnt 13, stanniocalcin, other renin-angiotensin system components, and vascular endothelial-derived growth factor, among others, have been shown to be mitochondria-localizing intracrines. The implications of this mitochondrial intracrine biology are discussed.

it has become apparent over recent years that a wide variety of peptide extracellular signaling molecules, such as hormones and growth factors, also act in the intracellular space after either internalization by target cells or retention by their cells of synthesis (44–60, 62). We have termed these factors intracrines and have developed proposals regarding their modes of actions. Intracrines are structurally diverse. Factors such as enzymes, hormonal fragments, and transcription factors are represented among the intracrines by virtue of their intra- and extracellular signaling activities (Table 1). While many intracrines traffick to the nucleus to directly or indirectly regulate transcription, others traffick to other intracellular locations such as mitochondria. We have suggested that just as intracrines participate in the regulation of cell growth and differentiation through the establishment of intracellular regulatory loops, they also participate in the regulation of metabolism as exemplified by intracrine enzymes (46).

Table 1.

Intracrines

Hormones, Cytokines	Growth Factors	DNA Binding Proteins	Enzymes	Other
Insulin	FGF (1,2,3,10)	Homeoproteins	Phosphoglucose isomerase/neuroleukin	Lactoferrin
Angiotensin II	Midkine	Amphoterin (HMGB1)	Renin/prorenin (aspartyl-protease)	Endogenous Opioids (Dynorphin)
				Galectins
Prolactin	VEGF	IL-33	PD-ECGF/thymidine phosphorylase	Tat
INF-β, -γ;	NGF		Granzyme A, B	Defensins
Interleukins	PDGF		PLA2-I	SHBG
PTHrP	Pleiotrophin		Lysyl-tRNA synthetase	Ribosomal protein S 19
Oxytocin	Proenkephalin		Thioredoxin	Pituitary adenylate cyclase-activating polypeptide
Leptin	IGF-1		Tyrosyl-tRNA synthetase	Endostatin
Growth hormone	Pigmented epithelium-derived factor (a serpin)		Pancreatic bile salt-dependent lipase	Periostin
Somatostatin	Maspin (a serpin)		Urokinase	Heat shock proteins
TRH	Schwannoma-derived growth factor		Trp-tRNA synthetase	PAI-2 (a serpin)
LHRH	Leukemia-inhibiting factor		Angiotensin-converting enzyme	Reelin
VIP	Macrophage colony-stimulating factor-1		AChE-R	PDCD5
Atrial natriuretic peptide	Hepatopoietin		Angiogenin	Thrombospondin-1
Gonadotropin	TGF-α			C-Peptide
Chronic gonadotropin	Hepatopoietin			STC
Angiotensin (1–7)	Heregulin			IGF BP-3,5,7
Endothelin	TGF-β			TCTP
Neuropeptide Y				Wnt 13
Erythropoietin				Osteopontin

PTHrP, parathroid hormone-related protein; TRH, thyrotropin-releasing hormone; LHRH, luteinizing hormone-releasing hormone; VIP, vasoactive intestinal polypeptide; HMGB1, high-mobility group protein B1; PD-ECGF, platelet-derived endothelial cell growth factor; PLA2-1, phospholipase A2-I; AChE-R, acetylcholinesterase; SHBG, sex hormone-binding globulin; PAI-2, plasminogen activator inhibitor-2; PDCD5, Programmed cell death 5; STC, stanniocalcin; BP, binding protein; TCTP, translationally controlled tumor protein. Ref. 28 and 60 document the correctness of entries..

Mitochondria play multiple roles indispensable for eukaryotic cell survival. Included among these functionalities are the generation of ATP though aerobic metabolism, the generation of reactive oxygen species (ROS), and the regulation of apoptosis. The institution of aerobic glycolysis, the Warburg effect, plays an important role in malignant cell biology (30). As generators of ROS, mitochondria sustain significant damage and must be removed from the cell through autophagy in a timely and appropriate fashion (38). Similarly, mitochondria must divide and partition themselves between daughter cells each time cell division occurs. It is therefore clear that mitochondrial biology is integral to the functioning of the eukaryotic cell. Early on, angiotensin II and transforming growth factors-β (TGF-β1) were shown to traffick to mitochondria, and subsequently other intracrines were similarly found in association with these organelles (Table 2) (14, 21, 24, 57, 63). Some of these factors also traffick to the nucleus or other intracellular sites. Recently, considerable interest has centered on the role of mitochondria in the determination of life span and the predisposition to degenerative disorders including cardiovascular disease (4, 9, 13). The generation of ROS and alterations in the concentrations of tricarboxy acid cycle intermediates have been suggested to play a role in these processes as has an interaction between mitochondrial function and the activation of hypoxia-inducible factors (HIFs) (73, 75, 76).

Table 2.

Mitochondrial intracrines

Angiotensin II

Angiotensinogen

Angiotensin-converting enzyme

Atrial natriuretic peptide

Growth Hormone

Renin

Stanniocalcin

TGF-β

Thioredoxin*

VEGF

Wnt 13b

^*See text for clarification.

Here we consider the possibility that intracrines participate in the regulation of mitochondrial function in health and disease, thereby exploring another dimension to their regulatory activities.

TGF-β

TGF-β are important fibrogenic cytokines with important roles in apoptosis and senescence (5, 24, 64, 68, 69, 74, 76). Liver stellate and Kupffer cells contain the large latent TGF-β complex consisting of the mature TGF-β protein linked noncovalently to the NH₂-terminal portion of the precursor protein, latency-associated peptide. This smaller complex is disulfide bonded to the latent TGF-β binding protein to create the large latent TGF-β moiety (64). Latent TGF-β is secreted and can be activated in the extracellular space. Moreover, it can be internalized by liver parenchymal cells that contain the complex but express no TGF-β message. As early as 1991, TGF-β was identified in the mitochondria of cardiac myocytes and hepatic cells and was also seen in association with the contractile filaments of cardiac myocytes and to a lesser extent with nuclei of hepatic stellate cells and myofibroblasts. Thus TGF-β is an intracrine with potentially important activity at the mitochondria. Of potential interest in this regard is the report that angiotensin II acting at nuclear angiotensin II type 1 (AT₁) receptor sites directly upregulates TFG-β1 transcription in renal cortical cells (32). This suggests the existence of an angiotensin II/TGF-β intracellular regulatory loop like those observed in the case of other intracrines (48, 54, 55, 58). Moreover, high glucose upregulates TGF-β in mesangial cells and stimulates extracellular matrix production. Angiotensin-converting enzyme inhibitors (ACEIs) do not reduce intracellular angiotensin II levels, TGF-β, or matrix production. A knockdown of cell angiotensinogen lowers intracellular angiotensin II, TGF-β, and inhibits matrix production (68). Thus the intracellular angiotensin II/TGF-β interaction is functionally relevant. Because both angiotensin II and TGF-β traffick to mitochondria, as well as the nucleus, and because angiotensin has important effects on mitochondrial number and function, the possibility that the two intracrines interact to regulate mitochondrial function through an intracellular regulatory loop should be considered.

TGF-β regulates the apoptosis of T lymphocytes but also, in the presence of serum, induces senescence in lung endothelial cells (5, 69). Mitochondria are very susceptible to damage from ROS and must be cleared through autophagy as damage accumulates. Depending on the conditions, TGF-β increases ROS generation and thereby induces senescence (76). At the same time, it can be noted that the hepatocytes of TGF-β knockout mice demonstrate an increased number of phenotypically normal mitochondria (74). This suggests that one function of TGF-β is to regulate the removal of mitochondria damaged in part secondary to the cytokine's stimulation of ROS. Alternatively, the increased mitochondrial numbers seen in the knockout model could represent a compensatory mechanism to offset less efficient mitochondrial function in the knockout animals.

Stanniocalcin

Stanniocalcin 1 (STC) is a secreted signaling protein that stimulates phosphate reabsorption in the renal tubules (65, 66, 77). It is synthesized in a variety of tissues including bone, prostate, renal tubules, ovary, and thyroid among others. Detectable levels of circulating stanniocalcin are found only during pregnancy and lactation. In tissues such as the renal tubules, selected cells express STC message and contain and STC protein, whereas nearby cells contain only protein. This finding, along with the uptake of STC by irradiated fibroblasts, indicates the internalization of STC by target cells (60). Also, STC receptors and STC are found not only on external cell membranes but also on mitochondria as well as on the nucleus. In renal tubular cells, more than 90% of cell-associated STC is associated with mitochondria. Collectively, these findings demonstrate that STC is an intracrine which trafficks to mitochondria and to a lesser extent to the nucleus (60, 65, 66, 77).

STC receptors have been detected on isolated mitochondria. STC itself can be detected in isolated mitochondria, and fractionation studies clearly indicate that it is associated with the inner mitochondrial matrix. Moreover, studies involving the addition of recombinant STC to submicrosomal particles indicated that STC produced a concentration-dependent stimulation of mitochondrial electron transfer beginning at a concentration of about 5 nM (66). STC appears to upregulate uncoupling protein 2 in macrophages and uncoupling protein 3 in cardiomyocytes, and there is evidence to indicate that STC does lead to uncoupling of the mitochondrial respiratory chain. In macrophages this uncoupling leads to less efficient ATP formation and a reduction in the generation of ROS (2, 16, 65, 66, 77). It will be recalled that uncoupling protein 1 is found in brown fat and is responsible for nonshivering thermogenesis. Uncoupling proteins 1 and 2 have little or no effect on thermogenesis but affect processes such as insulin secretion and macrophage function (2, 16).

It has been shown that mesenchymal stem cells apparently protect irradiated fibroblasts from apoptosis by secreting STC, which is internalized by the fibroblasts (60). Similarly, STC is constitutively expressed by terminally differentiated neurons. When a neural crest cell line was exposed to recombinant STC, the cells' uptake of inorganic phosphorus was increased. When the cells were exposed to increased extracellular calcium concentrations, STC was upregulated. The upregulation of STC by transfection in these cells enhanced their resistance to hypoxia and blunted intracellular calcium rises in the face of an inhibitor of endoplasmic reticulum Ca²⁺-ATPase. Consistent with this, STC is upregulated in neurons surrounding the infarction zone in ischemic brain injury in humans and rats. Because increased intracellular free calcium is an important mediator of ischemic neuronal death, it is possible that STC, a glycoprotein first identified as a regulator of calcium and phosphate metabolism in fish, protects neurons by blunting intracellular rises in free calcium (77).

Thus STC is an intracrine with important actions at the mitochondria. It is antiapoptotic and cell protective in several systems. Whether these effects are related to mitochondrial uncoupling with reduced ROS and ATP formation, the result of blunting the concentration and/or effects of intracellular calcium, direct effects on uncoupling proteins, or other mechanisms remains to be determined.

Growth Hormone

Growth hormone regulates growth and metabolism both by stimulating the secretion of insulin-like growth factor 1 (IGF-1) and by direct action at growth hormone receptors. Growth hormone is also internalized and thereafter trafficks to the nucleus and mitochondria. At the mitochondria, growth hormone binds to specific, high-affinity receptors and directly regulates mitochondrial respiration. More specifically, it decreases the activity of succinate dehydrogenase and cytochrome c oxidase in a specific and dose-dependent fashion. This effect can be contrasted with the effects of growth hormone on intact cells where it stimulates cytochrome c oxidase via an interaction with cell surface receptors. Presumably, intracellular growth hormone mitigates the stimulation of ROS produced by extracellular hormone (1, 67).

Angiotensin II and Renin-Angiotensin System Components

In the 1970s, tritiated angiotensin II, when administered to rodents, was shown to rapidly traffic to the nucleus and mitochondria, and specific, high-affinity angiotensin II receptors were detected on isolated mitochondria and nuclei; subsequently, angiotensin II immunoreactivity was detected, associated with the euchromatin and mitochondria of unmanipulated animals (17, 21, 61, 63). An emerging school of thought holds that dysfunctional mitochondria contribute to the pathophysiology of hypertension, cardiac failure, metabolic syndrome, obesity, diabetes mellitus, renal disease, atherosclerosis, and aging (11, 13, 14). Studies from various laboratories have demonstrated that the treatment of rodents with ACEIs or angiotensin II AT₁ receptor blockers (ARBs) prolongs life span and maintains mitochondrial number and morphology over time (4, 9). It has been shown that angiotensin II treatment of a variety of cells increases mitochondrial ROS production, decreases mitochondrial membrane potential, increases cellular superoxide production, and decreases nitric oxide bioavailability. The latter leads to decreased endothelium-dependent relaxation and a propensity to vasoconstriction. Angiotensin II-stimulated production of mitochondrial superoxide is dependent on NADPH oxidase activity, and animals lacking components of this enzyme are resistant to angiotensin-induced hypertension. Moreover, mitochondrial superoxide stimulates extramitochondrial NADPH oxidase activity in a feed-forward manner (13). Collectively, these observations suggest that angiotensin II might mediate pathological effects directly through an interaction with mitochondria.

It has been shown that ARB treatment improves kidney mitochondrial function in spontaneously hypertensive rats (SHRs) (12). When 4-mo male rats were treated with candesartan versus vehicle, candesartan lowered SHR systolic blood pressure (BP), proteinurea, and cortical glomerular area and increased creatinine clearance relative to vehicle treatment or to Wistar-Kyoto (WKY) rats. Moreover, SHRs were shown to have lower kidney mitochondrial membrane potential and nitric oxide synthase and cytochrome oxidase activities than WKY rats, the values of which were corrected by candesartan treatment. Candesartan was proposed, therefore, to preserve mitochondrial function. In later studies, this laboratory showed that while amlodipine reduced SHR BP to WKY levels, it was much less effective than losartan in correcting renal damage. This was termed the “mitochondrial antioxidant effect of losartan.” In another study, angiotensin II receptor blockade was shown to protect kidney mitochondria in streptozotocin-induced Type I diabetes (10). When 8-wk male Sprague-Dawley rats, rendered diabetic by streptozotocin injection, were treated for 4 mo with losartan or amlodipine, losartan was more effective in reversing renal lesions, plasma glucose, and proteinurea. Mitochondrial H₂O₂ and uncoupling protein 2 were also higher in amlodipine than losartan; cytochrome oxidase c activity and renal glutathione were lower in amlodipine than losartan. Amlodipine was actually slightly more effective in lowering BP, reducing systolic BP by an additional 5% over losartan. These measurements collectively support the proposal that AT₁ receptor blockade protects kidney mitochondria independent of BP.

In addition to the many kidney studies reporting an association between mitochondrial dysfunction and the renin-angiotensin system, several studies have reported changes in brain mitochondria related to hypertension. Others compared mitochondria from 12-wk SHRs with age-matched WKY rats (by two-dimensional electrophoresis followed by mass spectrometry and by Western blot analysis combined with sucrose-gradient ultracentrifugation/tandem mass spectrometry) and found previously unknown assembly defects in complexes I, III, IV, and V in the hypertensive rats (33). This represents further evidence that SHRs possess abnormal mitochondria which may contribute to the development of pathology. It further implicates specific members of the electron transport pathway in hypertension. In a more recent study, losartan was found to improve respiratory function and coenzyme Q content in brain mitochondria of young SHRs (71). Losartan was administered from the fourth to the ninth week of age and reduced BP and cardiac hypertrophy. In addition, losartan partially reversed the decline in brain weight-to-body weight ratio observed in SHRs. ATP production, which was also reduced in SHRs, increased with losartan treatment.

When administered to whole animals, angiotensin II increases ROS production and reduced energy generation. In some cases, it increases the levels of uncoupling proteins 1 and 3 and decreases the transcription of mitochondrial respiratory chain proteins. Angiotensin II infusion has also been shown to alter mitochondrial number, and this effect could be partially blocked by either AT₁ or AT₂ blockers (4, 9, 14, 35). Studies in cultured C2C12 cells showed that angiotensin II downregulated genes associated with mitochondrial biogenesis; this effect is mediated by the AT₂ receptor (35). In intact animals, the mechanism of angiotensin II modulation of mitochondrial number is more complex in that it appears that the peptide can also induce mitochondrial autophagy through the production of ROS. These effects on mitochondrial number assume immediate relevance when one recalls that a decreased skeletal muscle mitochondrial number is associated with prediabetic and diabetic subjects. This raises the possibility that ACEIs and ARBs improve glucose tolerance by maintaining the mitochondrial number (35). There is evidence that in some cases, intracellular angiotensin II can upregulate TGF-β, and therefore some of its mitochondrial actions, including the effects on the mitochondrial number, could be mediated by TGF-β (32). It has also been demonstrated that the ACEI, perindopril, reduced the elevated vascular endothelial growth factor (VEGF, an angiogenic nuclear/mitochondrial intracrine)-to-pigment epithelium-derived factor (PEDF, an anti-angiogenic intracrine) ratio found in diabetic rat retina, and the drug also reduced retinal damage. In cultured bovine retinal capillary endothelial cells, perindopril upregulated peroxisome proliferator-activated receptor-γ and uncoupling protein 3, reduced ROS production, and thereby lowered the VEGF-to-PEDF ratio (78).

Thus angiotensin II has important effects on mitochondrial function and on the mitochondrial number. The effectiveness of ARBs in mitochondrial rescue studies has been shown, in many cases to be independent of BP, suggesting that the classic AT₁ receptor signal transduction pathways may not be the only target of ARBs; ARBs may block not only cell surface signaling pathways but other events as well. We suggest that one of the alternative events could be an internalization and release of angiotensin II into cells where it may act directly upon the mitochondria. Endosomes are known to be leaky; their contents are frequently spilled into the cytoplasm (6, 7) This “endosomal escape” concept is, indeed, the basis for many mammalian DNA transfection technologies. ARBs may block angiotensin II action at the mitochondria by blocking angiotensin II release into the cytoplasm and, thus, access to the mitochondria.

Early studies clearly show direct angiotensin II effects on ATP production by aged mitochondria as well as other effects, but the full spectrum and the physiological implications of angiotensin mitochondrial action are unknown (21). Too little work has been done on isolated mitochondria to clarify this issue. However, the available data suggest the possibility that intracrine angiotensin II acts at the mitochondria to alter ROS generation and other organelle functions leading to vasoconstriction, hypertension, and vascular disease. Our recent development of transgenic animals expressing an intracellular angiotensin II fusion protein could be helpful in this regard (62), as could application of technologies designed to identify molecular binding partners for angiotensin II.

Also of interest is the observation in sheep that angiotensin II acts at the nucleus to stimulate the generation of ROS in the nucleus via an AT₁ mechanism. Angiotensin II stimulation of AT₂ nuclear membrane receptors upregulates nuclear nitric oxide production. The relative abundance of nuclear AT₁ and AT₂ receptors changes with age, and angiotensin (44, 45, 48–50, 52, 53) action at the nucleus modulates the effect of angiotensin II on nuclear ROS generation. These findings raise the possibility that parallel intracrine systems exist in the mitochondria and nuclei and modulate ROS activity. Arguably, this also emphasizes the importance of the endosymbiotic origin of the eukaryotic cell in determining modern cell biology (22, 23, 41).

Angiotensin-converting enzyme, renin, and angiotensinogen, all of which have been shown to be intracrines, have been found in adrenal mitochondria (42, 56). The renin appears to result from an alternative transcript lacking a signal sequence and is predicted to be synthesized in an active, as opposed to the proenzyme, form. This intracellular mitochondria-associated renin is markedly upregulated in anephric rats and participates in the maintenance of aldosterone synthesis in these rodents. It seems reasonable to suggest that a mitochondrial renin-angiotensin-aldosterone system exists in these cells and that intramitochondrial angiotensin II synthesis occurs along with any direct actions of these intracrine factors at that site (42).

Atrial Natriuretic Peptide

In regard to possible direct mitochondrial effects of intracrine renin-angiotensin system components, it is interesting to note that atrial natriuretic peptide (ANP) blunts aldosterone secretion. ANP is also an intracrine, and adrenal cortical cells are among those target cells that have been shown to internalize it. High-affinity binding sites for ANP have been identified on adrenal mitochondrial membranes, and intracellular ANP has been found in association with mitochondria (25, 36, 37). Other studies have shown that the direct application of recombinant ANP to renal mitochondria inhibited step 3 mitochondrial respiration and decreased the ADP-to-oxygen ratio; low molecular weight ANP also induced marked mitochondrial swelling (29). Collectively, these results indicate that ANP is an intracrine which likely acts directly at mitochondria, and these findings further support the notion that the ability of ANP to block aldosterone secretion is at least partially the result of a direct intracrine action at mitochondria.

Wnt 13

Wnt proteins are important mediators of development, cell proliferation, the establishment of polarity, and cell migration (60, 70). There are 19 members of the Wnt family. These are secreted and act after binding to the frizzle receptor and its coreceptors lipoprotein receptor-related proteins 5 and 6. Disheveled proteins serve as intracellular mediators of three intracellular signaling pathways. One member of the Wnt family, Wnt 13, acts in a more complex fashion. By using alternative promoters and alternative slicing, this gene generates three transcripts. In addition, two translation start sites are employed to generate additional isoforms, two of which, although resulting from two messages, lead to the same protein that trafficks to the nucleus. Another translation product produces a Wnt protein that trafficks to mitochondria. A third isoform is secreted. Thus Wnt 13 is a classical intracrine; in fact, it uses the multiplicity of methods for the formation of variants (alternative promoters, splice sites, and translation start sites) that are characteristic of archetypical intracrines.

Wnt 13 trafficking to the nucleus, but not to the mitochondria, is associated with mitochondrial fragmentation and apoptosis. Because Wnt 13 is expressed in endothelial cells, the distribution of Wnt translation products could have important effects in vascular biology/disease, as well as in development. However, the role of Wnt at the mitochondria is as yet unknown (60, 70).

VEGF

VEGF is an intracrine known to traffick to nucleus. It participates in intracellular intracrine regulatory loops that play a role in hematopoietic stem cell development, myeloma, breast cancer cell proliferation, and angiogenesis. VEGF has also been shown to localize at the mitochondria. However, there is no evidence indicating the nature of its actions, if any, at that organelle. Nonetheless, VEGF is a mitochondrial intracrine, and given its important actions in the vasculature, its likely mitochondrial actions deserve further investigation (18, 31, 57).

Additional Peptides/Proteins of Interest

Glucagon receptors have been identified on mitochondria, and glucagon administration to isolated mitochondria increases their ATP production in a substrate specific fashion (8, 26). However, there is no evidence to clearly indicate that glucagon trafficks to mitochondria, although this possibility must be considered likely. It is also of note that insulin is a major regulator of metabolism and trafficks to the nucleus. There is only very sketchy evidence, however, suggesting that insulin levels at the nucleus affect the metabolic state such as insulin resistance, and there are no reports of insulin trafficking to, or binding to, isolated mitochondria (55). Given the biological and medical importance of insulin and glucagon action, the available results suggest the need for more detailed studies of possible glucagon and insulin intracrine action at the nucleus and mitochondria (8, 26). The presence of mitochondrial binding sites for glucagon and the demonstration of direct glucagon action at the mitochondria strongly suggest that glucagon is an intracrine that acts at the mitochondria.

Corticotropin-releasing factor (CRF) is a hypothalamic hormone. CRF receptors have been identified on mitochondria, but CRF trafficking to the mitochondria and CRF direct effects at the mitochondria have not been shown (27).

Thioredoxins (TRXs) in humans are a family of very similar proteins that, among other functions, serve as free radical scavengers (19, 20, 39, 40, 43, 72). Two members of this family, TRX-1 and TRX-2, each expressed in several isoforms (splice variants), are of interest here. They are the products of two genes: one encodes a secreted form of the compound, TRX-1, which, among other actions, can stimulate the growth of target cells, trafficks to the nucleus in its cell of synthesis to regulate transcription, and, in addition, can be internalized by target cells; the second gene encodes TRX-2, a TRX that is targeted to the mitochondria (39). Mitochondrial TRX subserves important functions at the mitochondria as evidenced by the fact that a knockdown of this moiety leads to intrauterine death secondary to severe apoptosis occurring just at the time of mitochondrial development (39, 40, 72). The secreted form of TRX is an intracrine; it manifests type I intracrine action because it is both secreted to serve as a signaling molecule and trafficks to the nucleus in its cell of synthesis to regulate transcription (60). Arguably, the mitochondrial form can be considered an intracrine because, although it is not the product of the same gene as the intracrine form, it does appear to be the product of a related gene, the related genes likely formed by gene duplication (50).

DISCUSSION

The intracrine view holds that, in part because of the origin of the eukaryotic cell as an endosymbiote, it is not surprising to find intracellular signaling molecules serving extracellular signaling functions as well. Indeed, from the point of view of an organelle such as the nucleus or mitochondria, the cytosol is arguably an “extracellular” space (49). Moreover, the intracrine hypothesis holds that there are three primary modes of intracrine action: intracellular signaling to coordinate cell growth with ribosomal function and cellular metabolism, the establishment of intracellular feed-forward loops designed to alter the cell state and therefore achieve cell differentiation or memory, and extracellular signaling to coordinate these process on the tissue and organism level. The role of intracrine biology, including intracellular regulatory loops, has been reviewed extensively elsewhere (44–52, 54–60). The observations reviewed in this article make clear that the mitochondrion is one of the intracellular targets of more than a few intracrine factors. This then raises the question of whether the principle of intracrine coordination of cell function through the use of finite gain positive feedback loops that has been shown for intracrine nuclear action finds parallels in intracrine action at mitochondria. This is a difficult area to address because little experimental work has been directed toward this area and because of the complexity of the various ways mitochondrial function can potentially feed forward to establish regulatory loops. For example, HIF-1 and -2 regulate the response to hypoxia, including the regulation of VEGF and erythropoietin. HIFs also regulate mitochondrial respiration and oxidative stress. Mitochondrial metabolism and oxidative stress also regulate the activation of HIFs (73). Thus a feed-forward metabolic loop exists. Whether intracrine factors affect this loop or create their own loops is unclear. Similarly, an overexpression of TRX-2, the mitochondria-localizing form of the protein, decreases total and mitochondrial ROS in the aortas of mice infused with angiotensin II. TRX-2 also has antiapoptotic effects and protects against cardiac hypertrophy and other abnormalities in angiotensin-infused animals (19, 39, 72). It has been suggested that angiotensin II activates NADPH oxidase, which in turn leads to elevated mitochondrial ROS, which appears able to further upregulate NADPH in a feed-forward fashion. It has been suggested that TRX-2 may act by interrupting this loop (13, 19, 72). If correct, this schema would represent an intracrine/metabolic loop analogous to the feed-forward transcriptional loops of intracrine action at the nucleus. Because many metabolic factors, including Krebs cycle intermediates, can serve as intracellular signaling molecules (for example, pyruvate, succinate, and fumarate, all of which affect gene transcription), these loops could become quite complex (3, 73, 75).

Intracrine action at the mitochondria, in general, affects one or more of several parameters, including respiratory chain function, ATP and ROS generation, apoptosis/senescence, and mitochondrial number. The effects on life span are, in some cases, associated with these effects of intracrines on mitochondrial biology, but it is not clear that intracrine action at the mitochondria is the driver of these longevity changes (9, 24, 34, 60, 64).

One phenomenon that may be particularly relevant in this regard is the regulation of mitochondrial number. Because of exposure to ROS, mitochondria sustain a high rate of oxidative damage, and damaged mitochondria must be destroyed and replaced. The importance of the proper and timely elimination of damaged mitochondria is demonstrated by studies of mice lacking the essential autophagy-regulating gene Atg7 in the hematopoetic system. These animals develop severe anemia. Their erythrocytes accumulate damaged mitochondria, and cell death results. A similar phenomenon occurs in the lymphocytic lineage (38). Normal cell division also requires a division of mitochondria in a coordinated fashion. Thus the regulation of mitochondrial biogenesis and autophagy is of great importance to cellular homeostasis. It is of note that the intracrines insulin-like growth factor 1 (IGF-1) and neuregulin can synergize to increase mitochondrial biogenesis acting through both ERK and phosphatidylinositol 3-kinase pathways, whereas angiotensin II and TGF-β1 can in some circumstances lead to a decrease in the mitochondrial content of target tissues (15, 35, 74). Whether intracellular actions of these peptides play any role in this regulation of mitochondrial number is unknown. In any case, a detailed understanding of mitochondrial biogenesis is likely to lead to an improved understanding of a variety of disorders including neurodegenerative diseases. For example, autosomal recessive Parkinson's disease is associated in many cases with a loss-of-function mutation in parkin. Parkin encodes a ubiquitin ligase, which also regulates mitochondrial homeostasis, likely by ubiquitizining a protein on the surface of damaged mitochondria and thereby marking them for autophagy. Loss of parkin function results in the accumulation of damaged, elongated mitochondria, and this could play a role in the pathogenesis of the disease (79).

In conclusion, there is good evidence for intracrine action at mitochondria. While the mechanisms of intracrine action at mitochondria remain to be worked out in detail, it is very likely that this intracrine biology is important in a wide variety of diseases.

GRANTS

This study was supported by the Ochsner Foundation and by National Heart, Lung, and Blood Institute Grant RO1-HL-072795.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).