Aging is associated with shifts in autocrine and paracrine pathways in the cardiac vasculature that may contribute to the risk of cardiovascular disease in older persons. To elucidate the molecular basis of these changes in vivo, phage-display biopanning of 3- and 18-mo-old mouse hearts was performed that identified peptide epitopes with homology to brain-derived neurotrophic factor (BDNF) in old but not young phage pools. Quantification of cardiac phage binding by titration and immunostaining after injection with BDNF-like phage identified a twofold increased density of the BDNF receptor, truncated Trk B, in the aging hearts. Studies focused on the receptor ligand using a rat model of transient myocardial ischemia revealed increases in cardiac BDNF associated with local mononuclear infiltrates in 24- but not 4-mo-old rats. To investigate these changes, both 4- and 24-mo-old rat hearts were treated with intramyocardial injections of BDNF (or PBS control), demonstrating significant inflammatory increases with activated macrophage (ED1+) in BDNF-treated aging hearts compared with aging controls and similarly treated young hearts. Additional studies with permanent coronary occlusion following intramyocardial growth factor pretreatment revealed that BDNF significantly increased the extent of myocardial injury in older rat hearts (BDNF 35 ± 10% vs. PBS 16.2 ± 7.9% left ventricular injury; P < 0.05) without affecting younger hearts (BDNF 15 ± 5.1% vs. PBS 14.5 ± 6.0% left ventricular injury). Overall, these studies suggest that age-associated changes in BDNF-Trk B pathways may predispose the aging heart to increased injury after acute myocardial infarction and potentially contribute to the enhanced severity of cardiovascular disease in older individuals.
- brain-derived neurotrophic factor
- functional proteomics
- Trk B
cardiovascular disease is the leading cause of morbidity and mortality in older individuals (2). Presently, myocardial infarction has a significantly worse prognosis in older persons, with higher mortality and complication rates, than in younger individuals (3, 39, 42), suggesting that senescent changes in the cardiovascular system may predispose older individuals to increased cardiac pathology.
The aging heart has significant changes in vascular function, with diminished angiogenic capacity, endothelial dysfunction, and increased potential for prothrombotic activity (for review see Ref. 49). Previously we described (8) the results of an in vivo phage-display cardiac biopanning study aimed at elucidating the mechanisms governing the dysregulation in PDGF-B-mediated cardioprotection in the aging rodent heart. Specifically, these studies identified an age-associated loss in a subpopulation of TNF receptor 1 microvascular endothelial cells that underlies the impairment in TNF-α-mediated PDGF-B expression, which promotes cardiac angiogenesis and suppresses apoptosis to decrease myocardial injury in rodent models of myocardial infarction (8, 15, 50). Additional research based on these biopanning studies resulted in the identification of an age-related increase in TNF-α-induced reactive oxygen species and lipid peroxidation as a potential mechanism contributing to the dysregulation and loss of the TNF receptor pathways in the older heart (8, 17).
Because of the multifactoral effects of aging on the cardiovascular system, we sought to identify additional cardiac vascular receptor-mediated pathways that may govern the age-related changes in cardiac phenotype and function. Here we report the extension of phage-display studies that identified an increase in the brain-derived neurotrophic factor (BDNF) receptor, Trk B, in the aging rodent heart. Moreover, functional studies revealed that age-associated alterations in cardiac BDNF-mediated pathways can enhance inflammation and increase myocardial injury after myocardial infarction in the aging heart.
Phage-display peptide library in vivo cardiac vascular biopanning.
The age-associated changes in cardiac microvascular surface receptors were probed by in vivo phage-display biopanning with a cyclic peptide pSKAN phagemid library (6 amino acid variable, ∼107 total complexity; Mo Bi Tec) as previously described (8, 17). Young female adult (3 mo old) and aging (18 mo old) C57BL/6 mice (National Institute on Aging Rodent Colony) were anesthetized with Avertin (0.015 ml/g) and injected with phage peptide library phage [1012 colony-forming units (CFU)/200 μl PBS] via their tail veins. Four minutes after injection the mice were killed and the hearts were rapidly explanted. The phage were then recovered with WK6λmutS Escherichia coli. Age-specific phage pools were amplified and titrated for two additional rounds of biopanning enrichment. The phagemid DNAs of the resultant clones (n ≥ 100 3- and 18-mo-old heart pools) were sequenced, and translated motifs were analyzed for homology to known cytokines (FASTA3), as determined by the first homologous mammalian sequence identified with E value <1 as previously described in the identification of the age-associated changes in TNF receptor pathways (8): two clones isolated from the old, but not young, heart pools [ARRGQA (ψO40) and GRRFIR (ψO145)] revealed homology to 1BNDA, peptide sequences BDNF5–9 and BDNF100–106, respectively. In addition, to probe the structural relevance of the ψO40 motif to exposed potential receptor binding domains of BDNF, the region of homology was mapped in a tertiary model and labeled by Cn3D3.0 software as previously described (8).
Individual phage clone in vivo cardiac vascular binding.
To confirm the age-associated differential cardiac binding capacity of the BDNF-like phage, injections with ψO40 phage clone were performed. ψO40 (1012 CFU in 200 μl PBS) was injected into both 3- and 18-mo-old mice as described above (n = 6/group). Titration of the phage incorporated into the hearts was used to quantify the ψO40 binding in the young and older hearts (n = 3 each). The phage clones were recovered from the explanted hearts with WK6λmutS E. coli, which were then titrated by serial dilution to measure CFU.
Protein analysis of cardiac BDNF-Trk B.
To investigate the potential age-related changes in BDNF-Trk B pathways, Western blotting for BDNF and Trk B was performed on protein isolated from 3- and 18-mo-old murine hearts. Briefly, antibodies immunoreactive with murine BDNF (sc-546, Santa Cruz) and the extracellular domain of Trk B (sc-8316, Santa Cruz) were used for Western blots of 10% SDS-PAGE gels with 100 μg of protein samples from 3- and 18-mo-old hearts (n = 3 each). Immunostains of cardiac sections were performed with antibodies to both the extracellular and COOH-terminal truncated domains of Trk B (sc-8316 and sc-8312, Santa Cruz), which were visualized with diaminobenzidine (DAB) and an avidin-biotin complex (ABC) kit (Vector). To study the potential interaction of the ψO40 phage with Trk B, additional hearts (n = 3/group) were harvested after ψO40 biopanning (described above) for sectioning and immunostaining with biotin-labeled monoclonal antibody to phage coat protein pIII (pIII, Mo Bi Tec) visualized with a FITC-streptavidin secondary conjugate and costained with the extracellular Trk B antibody with a Texas red secondary antibody.
Induction of transient cardiac ischemia.
Transient cardiac ischemia (hypoxia) was induced in 4- and 24-mo-old F344 female rats (National Institute on Aging Rodent Colony) as a model of clinical myocardial ischemia (23) to investigate the potential alterations in BDNF pathways. Because of the marked heterogeneity of the coronary anatomy of the murine heart, which requires intramyocardial sutures to ensure ligation of bridging arteries (1, 43), we elected to use the rat cardiac model system, which has a more consistent coronary anatomy for myocardial ischemia and infarction studies (8, 15, 50). Specifically, sets of 4- and 24-mo-old rats were randomized to transient induction of cardiac ischemia and sham operation (n = 5/set). After anesthesia with ketamine (90 mg/kg ip) and xylazine (4 mg/kg ip), a tracheotomy was performed and the animal was intubated and ventilated (Harvard Rodent Ventilator model 683) with room air. A left intercostal thoracotomy was performed at the fourth and fifth intercostal spaces. On identification of the left anterior descending artery (LAD), a 7-0 suture was secured ∼2.0 mm below the level of the tip of the normally positioned left atrium, and four rounds of 10-min ischemia and reperfusion were induced. After the conclusion of the fourth reperfusion the suture was removed, the chest wall was closed, and the animal was removed from the ventilator and placed under a heat lamp to recover. Sham-operated rats received similar treatment but without coronary ligation. Rats were euthanized 24 h after injection, and hearts were excised, fixed, sectioned for immunohistochemical staining with rabbit antibodies directed to BDNF (sc-546, Santa Cruz), and visualized with the ABC-DAB system and costaining with Giemsa. The number of immunopositive cells was quantified in sections at the midpapillary level of each heart, and the number of stained luminal structures in a total of six high-power fields (×40 magnification) per section was analyzed in a blinded fashion as previously described (8, 15).
In vivo response to BDNF in young and old rat hearts.
To probe the functional relationship between changes in BDNF-Trk B pathways and cardiac inflammation, additional sets of 4- and 24-mo old F344 rats (n = 3/set) received intramyocardial injections of BDNF and the extent of changes in cardiac infiltrates was assessed (8, 15). Briefly, the rats were anesthetized and underwent left intercostal thoracotomy, and the LAD was identified. On the basis of previous rat in vivo injection studies with supraphysiological, nontoxic concentrations of BDNF injected into retinal tissue (48) and intact brains (4, 22), 1 μg of recombinant human (rh)BDNF (248-BD, R&D Systems) in 50 μl of PBS or PBS alone was injected through a 30-gauge needle with a 250-μl Hamilton syringe in two injections (25 μl/injection, 2 mm apart) in the mid-left ventricular anterior wall of the rat hearts. The chest wall was then closed, the lungs were inflated, the rat was extubated, and the tracheotomy was closed. The following day the rats were killed, and the hearts were excised, fixed in 4% paraformaldehyde, and embedded in paraffin. Sections were stained with hematoxylin and eosin, immunostained with mouse anti-rat RM1, ED2, and ED1 (Research Diagnostics), and visualized with an ABC kit with DAB. Staining was quantified in the anterior left ventricular wall at the level of the midpapillary muscles from each heart, as previously described (Refs. 8, 15; 10 high-power fields/heart). Two investigators performed quantification independently in a blinded fashion.
Myocardial infarction studies.
To study the potential age-dependent effects of BDNF in a cardiac injury model, a rat myocardial infarction model was used in which the level of coronary ligation was age adjusted to provide similar-size myocardial infarctions in young and old rats, as previously described (8, 15). Sets of 4- and 24-mo-old F344 rats received intramyocardial injections of rhBDNF (1 μg) or PBS (n = 5 each) as described above. The next day the rats were reanesthetized, the heart was exposed, and the LAD was ligated directly below (4 mo old) or 2 mm below (24 mo old; to produce a similar-size myocardial infarction to the younger rats) the left atrial appendage with 8-0 nylon sutures. Pallor and regional wall motion abnormality of the left ventricle confirmed the occlusion. The chest wall was closed, and after recovery the rats were returned to the animal facility for 14 days. At the termination of the experiment the rats were imaged by transesophageal echocardiography as previously described (53) and then killed, and the hearts were explanted. The extent of myocardial infarction measured at the level of the midpapillary heart muscles was scored by Masson's trichrome staining, and the images were analyzed in a blinded fashion with ImageJ 1.22 software (NIH Image) (36, 47). Infarction size with linear approximations to account for area gaps in histology was expressed as a percentage of the total left ventricle myocardial area as previously described (8, 15).
Comparisons of categorical variables were conducted with the binomial distribution test or Fisher's exact test, as appropriate. Those between nonnormally distributed continuous variables employed the Wilcoxon rank-sum test. A P value <0.05 was considered statistically significant.
In vivo cardiac homing of BDNF-like phage to the aging heart.
In vivo cardiac biopanning revealed clones with homologies to BDNF in the older but not younger cardiac-homing phage clones (2/100 vs. 0/101, P < 0.05), including ψO40, which is homologous to the amino-terminal domain of BDNF that mediates high-affinity ligand interactions with Trk B (34) (Fig. 1, A and B). Titering of cardiac eluted phage confirmed the increased cardiac homing in the older murine heart (Fig. 1C), suggesting that BDNF binding sites/receptors may be increased in the aging cardiac microvasculature. Western blots revealed that BDNF levels were comparable in the young and old hearts. Trk B, however, was greater in the aging heart samples. Notably, the isoform of Trk B expressed was the truncated receptor, with a known size of ∼95 kDa (20) (Fig. 1D). The differential expression of Trk B was confirmed by immunostaining of cardiac sections from 3- and 18-mo-old mice (Fig. 1, E and F). Moreover, immunofluorescent staining of the hearts after ψO40 phage clone injection demonstrated that the age-associated increase in the BDNF-like phage colocalized with Trk B in the cardiac microvasculature (Fig. 1G).
Age-associated BDNF-mediated cardiac inflammation.
On the basis of previous reports demonstrating that BDNF is induced in the ischemic (hypoxic) heart (25), a rat model of transient coronary occlusion was used to study potential changes in BDNF patterns in the aging heart. Immunostains revealed a more extensive staining of BDNF in the untreated aging hearts compared with young controls, corresponding to the age-associated increased density of Trk B in the heart. After transient episodes of cardiac ischemia, BDNF staining was also higher in the aging hearts compared with similarly treated young rats or age-matched sham-operated control rats (Fig. 2). Moreover, this age-associated increase in BDNF was associated with a mononuclear infiltrate in the old cardiac myocardium that was not observed in the young hearts (Fig. 3).
To probe the relationship between BDNF and cardiac inflammation, additional rats were treated with intramyocardial injections of BDNF. Histological analysis of these hearts revealed that BDNF did not significantly increase the cardiac mononuclear infiltrate in either the young or the old hearts (Fig. 3). Investigations of the potential age-related differences in the characteristics of these cells by immunostaining for different macrophage surface antigens revealed that the number of cells staining for RM-1, a cell surface antigen found on macrophage in fetal tissue (24, 45), in the cardiac sections was slightly, though not significantly, higher in the young hearts (Fig. 4). In contrast, ED2, a rat homolog of CD163 that is associated with early stages of macrophage activation (29, 41), was significantly greater in both the BDNF- and control-treated aging hearts compared with similarly treated young hearts. Notably, BDNF injections resulted in an increase in ED1, a rat homolog of CD68 (46) that is present on activated/inflammatory macrophages (11, 12), with higher concentrations in the BDNF-treated older hearts.
In vivo BDNF function in myocardial infarction in the aging heart.
To investigate the impact of the age-associated changes in BDNF-mediated pathways in an experimental model of myocardial infarction, additional 4- and 24-mo-old rats were treated with BDNF or PBS control followed by permanent coronary ligation. In the young hearts, intramyocardial injection of BDNF did not alter myocardial infarction size measured 2 wk after acute coronary occlusion compared with PBS controls (Fig. 5, A and B). In the old rats, BDNF was deleterious, with significantly more extensive myocardial injury compared with hearts injected with control vehicle. Moreover, the old hearts treated with BDNF developed a persistent inflammatory infiltration (Fig. 5C) associated with aneurysmal changes in left ventricular wall segments as observed by echocardiography (Fig. 5D), suggesting that the age-associated increase in Trk B and altered BDNF function may contribute to the severity of cardiovascular pathophysiology in the aging heart.
In the present report we describe the age-associated changes in BDNF-TrkB pathways that can result in enhanced pathology in the rodent heart. Specifically, in vivo phage biopanning studies revealed that age-related increases in cardiac microvascular truncated Trk B are associated with shifts in BDNF-mediated inflammatory responses in old hearts. Moreover, delivery of BDNF led to a senescent increase in myocardial injury after acute coronary occlusion, suggesting that the alterations in Trk B pathways in the aging heart may contribute to the more severe cardiac pathology observed in older persons with cardiovascular disease.
Previous studies have revealed the importance of BDNF in the development of the cardiac vasculature, promoting the stabilization of intramyocardial vessels (13). Moreover, although its actions in the adult heart have yet to be fully elucidated, recent clinical studies have shown that BDNF is associated with coronary atherosclerosis and unstable angina (19), and experimental studies have revealed that BDNF may be involved in ischemia-reperfusion cardiac injury (25). These findings, together with the increase in Trk B-positive cells in the senescent cardiac microvasculature, suggest that shifts in BDNF-mediated pathways may predispose the aging heart to increased vascular pathology.
The age-associated increase in Trk B is linked to an age-associated increase in inflammatory pathways and a significant increase in myocardial injury after coronary occlusion. Specifically, immunostains revealed an age-associated increase in ED2 staining, which is similar to the age-related increased number of perivascular ED2+ macrophages reported in rat brain (32). The BDNF-mediated increase in ED1 staining was previously shown to correlate with inflammatory activation and rejection in cardiac transplant studies (26). Moreover, the inverse relationship between ED1 and RM-1 is consistent with patterning of activated macrophages in experimental rat models of pulmonary fibrosis (27). Furthermore, previous studies in the aging rat thymus have reported a link between increased ED1+ macrophages and Trk B (21), suggesting that macrophage activation in the heart may be directly related to the age-associated increase in BDNF-Trk B pathways.
Previous studies have demonstrated an association between BDNF-Trk B signaling and vascular inflammation (14). Specifically, the paracrine actions of truncated Trk B receptor have been suggested to be involved in the perineural response to peripheral inflammation (30). Moreover, the combination of the age-associated alterations in macrophages (33) that may also express Trk B (40) and the potential recruitment of bone marrow-derived Trk B+ cells (28) may act with the age-associated increase in truncated Trk B in the cardiac microvasculature to govern the BDNF-mediated inflammatory induction in the older hearts. Overall, the age-associated changes in BDNF-Trk B pathways may enhance vascular inflammation to increase the severity of cardiovascular pathophysiology in older persons and may provide a functional link to recent clinical studies that have demonstrated the role of inflammation as a cardiovascular risk factor in the geriatric population (9).
Mechanistically, the increase in truncated Trk B and the age-associated increase in immunomodulation may be related to the physiological requirements of the aging cardiac vasculature. Previous studies have demonstrated the importance of inflammatory pathways in the induction of angiogenesis in the heart (31). These findings, in conjunction with the role of age-associated impairment in cardiac angiogenic pathways (8, 16, 50), suggest that the shift in BDNF-Trk B pathways could be a compensatory response to augment angiogenic function in the aging heart. We recently demonstrated (8, 52) that shifts in TNF-α pathways underlie the loss in proangiogenic/antiapoptotic cytokine-mediated function in the older rodent heart. Thus the interrelationship of TNF-α and BDNF in endothelial cells (6) and the antiapoptotic actions of BDNF on Trk B cells (13) may play an important role in maintaining vascular homeostasis in the aging heart. Indeed, the changes in BDNF-Trk B pathways could be related to the age-related increases in TNF-α (7, 10) that can modulate induction of BDNF in monocytes (44).
In light of the multiple and complex interactions that may influence the activity of the BDNF function in the aging heart, we postulate that additional in vivo studies, including investigations with the ψO40 phage peptide, may facilitate the elucidation of the specific molecular and cellular mechanisms that contribute to the proinflammatory predisposition in the aging heart. Indeed, previous studies using peptides encoding sequences of BDNF have revealed both antagonist and partial agonist function in BDNF-Trk B interactions (37, 38), suggesting that similar approaches based on the structure of ψO40 could modulate the function of BDNF pathways. We anticipate that the BDNF phage-based peptides may be useful in probing the specific actions of truncated Trk B, which is known to compete with the full-length receptor (18, 51) but can also signal directly through non-tyrosine kinase signaling cascades (5, 35). These peptides may be useful in assessing the role of Trk B pathways in aging as related to hormonal signaling and other pathways that can modulate the biology of aging in the vasculature. Furthermore, we project that therapeutic approaches to counter the senescent immune activation induced by BDNF may exploit the biophysical interactions of ψO40 with Trk B as a potential means of inhibiting BDNF binding to Trk B in the aging heart.
This work was supported by grants to J. M. Edelberg from the American Federation for Aging Research-Beeson Physician Faculty Scholar in Aging Research and National Institutes of Health Grants AG-20320, AG-20918, and HL-67839.
Article published online before print. See web site for date of publication (http://physiolgenomics.physiology.org).
Address for reprint requests and other correspondence: J. M. Edelberg, Weill Medical College of Cornell Univ., 520 East 70th St., Box 161, New York, NY 10021 (e-mail:)
- Copyright © 2006 the American Physiological Society