Inflammation,Stem Cells,and the Aging Hypothalamus

Introduction

The hypothalamus is a small almond-shaped structure in base of the brain comprising several subregions or clusters of neurons called nuclei that control many homeostatic functions including circadian rhythm, body temperature, hunger, thirst, attachment behaviors, sleep, and fatigue/wakefulness. ¹ The hypothalamus plays a central role in neuroendocrine function acting as a nexus between the nervous and endocrine systems through releasing hormones manufactured and released upstream from the pituitary gland. Because of its central role in homeostasis, the hypothalamus has long been hypothesized to be a master “regulator” of aging (the so-called neuroendocrine theories of aging), ² and recent evidence suggests that indeed aging-associated molecular and cellular changes in the hypothalamus impact systemic homeostasis and life span in mammals. ^{3 –5} Regardless of theoretical questions surrounding the existence of “master regulators of aging,” the key integrative role of the hypothalamus in maintaining systemic homeostasis identifies it as a potential key point of organism-wide physiological loss of function.

The most compelling reasons to believe that the hypothalamus plays a key role in aging derive from two reports, whereby genetic manipulation of the hypothalamus results in extended life span and delayed aging-associated dysfunction. Mice that over express SIRT1 in the brain (BRASTO, brain-specific Sirt1-over expressing transgenic mice) have a life span extension of 16% for females and 9% for males. SIRT1 increases neural activity in the dorsomedial hypothalamic (DMH) and lateral hypothalamic (LH) nuclei of the hypothalamus, resulting in increased Ox2r expression, a G-protein receptor that is involved in feeding behavior. ⁵ Do these findings suffer from any potential artifacts? Food intake was monitored to alleviate calorie restriction artifacts, however, the BRASTO mice use the mouse prion (prp) promoter to express SIRT1 and prp supports SIRT1 expression in the kidney and heart as well as in the brain of the transgenic mice. Moreover, the prp promoter is known to be transcribed in immune cells, all of which may complicate interpretation of these data.

In another provocative study, increased nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) activity was observed with increasing age in the mediobasal region of the hypothalamus (MBH). Ectopic expression of an inhibitor of proinflammatory NF-κB/IκκB signaling in the MBH region of the hypothalamus increased life span in male mice. A dominantly acting IκκBα driven by CD11b a microglial-specific promoter using lentiviral vectors injected into the MBH region of the hypothalamus increased life span by 23% in male mice and decreased aging-related dysfunction in the nervous system and skeletal muscles: increasing neurogenesis, skin thickness, bone density, while decreasing collagen cross-linking. Furthermore, increased levels of NF-κB reduced expression of endocrine factor gonadotropin-releasing hormone (GnRH) by twofold. GnRH treatment attenuated the aging-associated phenotypes. ³ One potential caveat is that female mice were not studied in the life span extension studies, but responded similarly phenotypically.

Neural Stem Cells in the Hypothalamus Affect the Rate of Systemic Aging by Secretion of Exosomal miRNAs

In a potentially paradigm-shifting follow-up study to work that showed that inhibition of NF-κB-mediated neuroinflammation in the hypothalamus results in antiaging effects, Zhang et al. demonstrate that neural stem cells (NSCs) in the hypothalamus (htNSCs) are lost during aging, but can be preserved by genetic modification to withstand an inflammatory microenvironment, resulting in a reduction of the aging phenotype and life span extension. Of particular significance is that these htNSCs secrete factors and miRNA-containing exosomes that reduce aging-associated dysfunction, ⁴ and so have an effector cell function independent of their ability to differentiate into neurons or neuroendocrine cells. These findings extend our understanding of what stem cells are capable and emphasize the key role that their dysfunction or loss may play in aging.

First, Zhang et al. identified NSCs in the third ventricle wall of the MBH region of the hypothalamus in mice by observing coexpression of NSC biomarkers Sox2 and BMI1, as well as other biomarkers that are often expressed in NSCs: nestin, Musashi1, and Cxcr4. All of these biomarkers decreased with age, suggesting that few htNSCs remain in old mice. BrdU incorporation, which identifies cells that have undergone DNA replication, decreases to almost undetectable levels in old mice. ⁴

To determine the significance of htNSCs to normal physiology and aging, Zhang et al. infected cells in the MBH with engineered lentiviral vectors expressing the herpes simplex virus thymidine kinase (TK) gene driven off of the stem cell-specific Sox2 promoter. Treating the middle-aged animals (15 months) with ganciclovir, which is converted into a toxic product by TK, resulted in the loss of ∼70% of the SOX2-expressing htNSCs. Followed for more than a 3–4-month period, mice with ablated htNSCs displayed increased loss of novel object recognition, decreased treadmill performance, endurance, and coordination than untreated controls. Similar experiments using other engineered lentiviruses dependent on a different cytotoxic mechanism to kill htNSCs or the Bmi1 promoter instead of the Sox2 promoter yielded similar results ruling out major potential artifacts. Significantly, life span was affected by loss of htNSC as well: injecting younger mice (8-month-old) with the Sox2-TK lentiviruses resulted in about 10% shorter median and maximum life span. Interestingly, the total number of neurons in the MBH was not affected by the elimination of the majority of htNSCs. ⁴

Since it is much easier to reduce biological function to negatively affect health and life span, Zhang et al. sought to demonstrate the converse by preventing or slowing age-related dysfunction by injecting htNSCs from newborn mice that were genetically engineered to express green florescent protein (GFP) as a marker into the MBH of middle-aged mice. As might be expected from their earlier work, these neurons did not survive, most likely because of the inflammatory environment already present by middle age in the MBH. To overcome this problem, Zhang et al. engineered NF-κB inflammation-resistant neurons that constitutively express a dominantly acting IκBa, an inhibitor of NF-κB to protect the neurons from proinflammatory signals. Two months after injection, about 50% of these IκBα htNSCs survived. As controls, Zhang et al. injected GFP-labeled mesenchymal stem cells (MSCs) or astrocytes into the MBH of similarly middle-aged mice. Only mice that had received the inflammation resistant IκBα-htNSCs showed an approximate 10% increase in median and maximum life span, as well as improved locomotion, endurance, coordination, treadmill performance, novel recognition, and sociability 3–4 months postinjection, suggesting that htNSCs play a key role in maintaining systemic homeostasis and viability. ⁴

This is quite a significant increase in life span for a treatment started at late middle age (15 months), but is not as significant as the 23% median/20% maximum life span increase observed by the authors previously using a microglia conditional knockout of IκκB, ³ a cofactor necessary for NF-κb function. Perhaps this is not surprising as the IκκB knockout is expected to affect all cells exposed to the microenvironment including mature neurons and astrocytes as well as htNSCs. Another take-home lesson that will be familiar to biomedical scientists is that stem cells are not all alike, MSCs had no effect, although to be fair, neither the MSCs nor astrocytes were treated with equivalent IκBα anti-inflammatory lentiviruses and so would be expected to be subject to proinflammatory signals from the microglial cells in the MBH. The authors suggest that their survival, unlike the htNSCs, results from resistance to inflammation, ⁴ but there are numerous examples in the literature of inflammation causing dysfunction rather than cell death. In any case, a local increase in htNSCs in the MBH, a relatively small area of the brain, results in a significant antiaging effect.

Zhang et al. postulated that neurogenesis was unlikely to have made much difference in such a “short” period of 3–4 months, ⁴ although they provide no direct evidence for this. It would have been interesting to express an inducible inhibitor of cell cycle progression, for example, CDK inhibitor p21, or neural differentiation, to determine the role of cell proliferation and neurogenesis. Instead, they postulated that the endocrine role of the hypothalamus must be involved. Specifically, they hypothesized that exosomal miRNAs may play a key role. miRNAs have been shown to profoundly affect stem cell function and fate as well as affect aging and fibrosis in the heart. ^{6 –8} Furthermore, exosomes are known to be an efficient method of miRNA intercellular trafficking. ⁹ Zhang et al. used electron microscopy to observe a large number of multivesicular bodies, known to be precursors of exosomes, in the htNSCs in the MBH, but not in astrocytes from the same nucleus. Interestingly, cultured htNSCs produce many times more exosomes, with a 100-fold higher RNA content than cultured MSCs or astrocytes. Although hippocampal NSCs also produce significant quantities of exosomes, the miRNA content of exosomes from htNSCs is significantly higher with a different pattern of expression. Zhang et al. observed that miRNA levels of 20 miRNAs associated with htNSCs are substantially decreased in the cerebrospinal fluid (CSF) of middle-aged mice, a finding consistent with the earlier observations that htNSCs are a major source of miRNA-containing exosomes and that they decrease with age. To demonstrate that the htNSCs in the MBH generate the suspect miRNA-containing exosomes, mice were treated with lentiviruses bearing shRNA to knock down expression of Rab27a, a protein necessary for exosomal secretion. Young mice treated for 1 week showed substantially less miRNA in their CSF, and middle-aged mice injected with the Rab27 shRNA lentiviruses showed some mild impairments after 6 weeks, but clearly not equal to the effect of ablating htNSCs. ⁴

To demonstrate the functional consequences of exosomal delivery of miRNAs, htNSC-derived exosomes from cells expressing GFP to label the exosomes were injected into middle-aged mice using a cannula placed in the third ventricle of the hypothalamus for a period of 4 months three times a week. First, the exosomes helped to maintain htNSCs, and this was likely because of reduced hypothalamic inflammation, suggesting that the miRNA from the exosomes may have anti-inflammatory effects. Second, when exosomes are injected into mice whose htNSCs had been mostly ablated by a Bmi1-TK lentivirus and ganciclovir just before a 4-month course of exosome treatment, in other words the hypothalamus-based accelerated aging model created by this group, multiple phenotypes associated with loss of htNSCs and aging are ameliorated, including locomotion, coordination, treadmill capacity, novel object recognition, sociability, and spatial memory. Third, in normally aging mice treated with htNSC exosomes, these same aging-related phenotypes remain relatively stable for the 4-month period. ⁴ Altogether, these data demonstrate that htNSC-derived exosomal miRNAs as the authors have hypothesized or at least htNSC-derived exosomes play an important although partial role in dysfunction associated with aging.

It is important to mention that this group previously reported that levels of GnRH are diminished in the aging hypothalamus and that treatment with exogenous GnRH increases neurogenesis and also ameliorates the same aging-associated dysfunction in similar assays. ³ Of interest here is that Zhang et al. report that some of the htNSCs stain for GnRH and may be the source of at least some of the GnRH normally found in the hypothalamus. It would be interesting to determine how many of the aging effects resulting from the inflammation of the microglial cells hypothalamus can be reversed by the combination of GnRH and htNSC-derived miRNAs.

This work connects the aging-associated increase in inflammation of the MBH region of the hypothalamus with loss of htNSCs and provides data that htNSCs or their direct progeny play a significant role through a newly discovered aspect of their endocrine function, exosomal miRNAs in maintaining their own homeostasis, and opposing aging.

Medical Implications

The medical implications of the two articles from Cao's group are profound, because they suggest that a small region of the relatively small hippocampus can affect life span and age-associated loss of neurological and muscular function, as well as collagen cross-linking, skin thickness, and bone density, and that inflammation-resistant htNSCs, reduction of microglial inflammation, htNSC-derived exosomal miRNAs, or endocrine hormones such as GnRH can ameliorate many of the aging-associated phenotypes studied. It would interesting to know whether any of the other systemic functions controlled by the hippocampus, although by different nuclei or regions, such as the suprachiasmatic nucleus, which controls circadian rhythms, are affected as well. It is important to note that life span extension was only examined and observed for inflammation-resistant htNSCs in Zhang et al. ⁴ or inflammation-resistant hippocampal microglia in Zhang et al. ³ and not for exosomes or GnRH.

Because htNSC exosomes appear to require direct introduction into the hypothalamus, they cannot presently be considered a practical therapeutic modality. Introduction into CSF might allow transit to the hypothalamus, but would probably not be practical for antiaging therapies. However, it is interesting that blood plasma from human umbilical cords has been recently reported to revitalize hippocampus function in old mice. ¹⁰ Blood plasma contains a rich set of exosomes, although the effects were attributed to proteins such as Timp2, it is not difficult to believe that exosomes may be a beneficial component of cord blood plasma or even play a significant role in parabiosis experiments. Discovering drugs that stimulate htNSC exosome secretion or, better yet, protect htNSC from their inflammatory microenvironment or, even better still, reduce hippocampal microglia inflammation would be potentially useful to maintaining or extending both health span and life span.

Interestingly it was recently reported that acarbose, 17-α-estradiol (17αE2), and nordihydroguaiaretic acid reduce hypothalamic inflammation and extend life span in male mice. These three drugs reduced several proinflammatory parameters, including TNF-α secretion as well as the number of activated Iba-1-positive microglia and GFAP-positive astrocytes. Interestingly these drugs had no effects on female mice. ¹¹ This raises a very interesting question about the sex specificity of the Zhang et al. life span extension data because they only used male mice in their studies. Significantly, caloric restriction significantly reduced hypothalamic inflammation in both sexes, and is known to extend life span in rodents as well many other animals. Perhaps a key target responsible for life span extension effects of caloric restriction (CR) is the hypothalamic microglia and htNSCs. Ironically, the hypothalamus may play a role in detecting and initiating CR-protective effects ¹² as well as contributing to obesity and glucose resistance in response to overnutrition. ¹³

The question remains how CR or these drugs reduce hypothalamic inflammation, which begs the question of the origin of hypothalamic inflammation with age in the first place. Proposed mechanisms include RNA stress and immune dysfunction because of immunosenescence. ¹⁴ Some evidence exists that decreased integrity of the blood brain barrier (BBB) with aging ¹⁵ may contribute to brain inflammation in general by allowing systemic factors to interact with brain tissue, and this may very well play a key role. ¹⁶ And what is the cause of the BBB breakdown? Some evidence exists for vascular cell senescence playing a role. So if this chain of reasoning is correct, will elimination of senescent cells prevent hypothalamic inflammation and downstream systemic effects? Does senescence play a role in hypothalamus inflammation, although senescent neurons and glial cells have been described in the literature? The answers are unknown at present.

Is there any connection between the htNSC data and data that suggest that enhanced brain-specific SIRT1 activity extends murine life span through increased orexin type 2 receptor expression in the DMH and LH regions of the hippocampus? The answer is unknown, but it would be surprising whether these research groups did not at least try preliminary experiments to find a connection.

Because the hippocampus plays an integrative, bidirectional role in homeostasis, it is a candidate for helping to synchronize various systemic aging clocks that have been described. For example, the Horvath DNA methylation clock ¹⁷ includes tissue-type independence, which would seem to require some mechanism of synchronization that could be provided by the hypothalamus. However, despite the existence of data obtained from the Horvath clock, it remains more than possible that aging of tissues proceeds at different rates. The Horvath clock is really just an integrated large set of aging biomarkers based on epigenetics. Perhaps the decrease of htNSC-derived exosomes carrying a defined set of miRNA may prove a useful, although difficult to obtain, alternative novel biomarker for aging.

Conclusion

An emerging connection between age-associated inflammation in the MBH region of the hypothalamus and loss of stem cells may generalize to other tissues, but given the central role of the hippocampus in homeostasis, its significance may be no where greater. That NSCs in the hippocampus may also act as effector cells that communicate through secreted factors and exosomal miRNA suggest that stem cell biology is more complex than so far appreciated. New proregenerative strategies that alter stem cell viability and the stem cell milieu are likely to be developed but remain for the immediate future on the other side of the adjacent possible.