Endostatin and kidney fibrosis in aging: a case for antagonistic pleiotropy?

Abstract

A recurring theme of a host of gerontologic studies conducted in either experimental animals or in humans is related to documenting the functional decline with age. We hypothesize that elevated circulating levels of a powerful antiangiogenic peptide, endostatin, represent one of the potent systemic causes for multiorgan microvascular rarefaction and functional decline due to fibrosis. It is possible that during the life span of an organism there is an accumulation of dormant transformed cells producing antiangiogenic substances (endostatin) that maintain the dormancy of such scattered malignant cells. The proof of this postulate cannot be obtained by physically documenting these scattered cells, and it rests exclusively on the detection of sequelae of shifted pro- and antiangiogenic balance toward the latter. Here we compared circulating levels of endostatin in young and aging mice of two different strains and showed that endostatin levels are elevated in the latter. Renal expression of endostatin increased ∼5.6-fold in aging animals. This was associated with microvascular rarefaction and progressive tubulointerstitial fibrosis. In parallel, the levels of sirtuins 1 and 3 were significantly suppressed in aging mice in conjunction with the expression of markers of senescence. Treating young mice with endostatin for 28 days showed delayed recovery of circulation after femoral artery ligation and reduced patency of renal microvasculature but no fibrosis. In conclusion, the findings are consistent with the hypothesis on elevation of endostatin levels and parallel microvascular rarefaction and induction of renal fibrosis in aging mice.

aging is associated with a gradual functional decline of all systems. Individual organs display signs of fibrosis expressed to varying degrees, from nil to moderate (2, 12, 29, 49). Ockham's razor-based approach would envisage a preferably uniform explanation for such a singularity of an effect in a plurality of affected sites. Indeed, theories linking aging to redox imbalance, telomere attrition, and sirtuin deficiency, among others, have a sufficiently broad footprint to be applicable to diverse organ involvement (4, 6, 19–20, 22, 41, 48). Moreover, mitochondrial dysfunction (28), uncapping of telomeres (32), and reactive oxygen species (ROS) (36) all are known contributors to vascular aging. Recently, Sahin and DePinho (42) proposed that mitochondrial dysfunction, the point of convergence of telomeric, oxidative stress, p53, and suppressed sirtuin-induced metabolic aberrations, may represent the key pathway leading toward aging-associated disorders. As attractive as this hypothesis is, the mechanistic route from mitochondrial dysfunction to tissue fibrosis, a hallmark of aging-associated pathologic findings, is not obvious or trivial.

Microvascular rarefaction is a permanent companion of fibrotic process occurring with a relative synchrony in all organs partly due to impaired angiogenesis in the elderly individuals (24, 26). This nearly universal finding has been linked to multiorgan fibrotic transformation, on the one hand, and, on the other, to a limited growth of some tumors in the elderly (39). Considering the universality of diminishing capillary density and impairment of angiogenesis, it is tempting to speculate that this phenomenon is governed by a systemic imbalance between pro- and antiangiogenic factors. Such an imbalance may be the consequence of reduced levels of proangiogenic factors like VEGF, as has been recently shown (13). Alternatively or in addition to that, impaired angiogenesis in the elderly individuals could be a manifestation of an enhanced production of antiangiogenic factors. Endostatin, the first and foremost natural angiogenic inhibitor, is a 20- to 22-kDa COOH-terminal fragment of collagen XVIII (35), a ubiquitous component of subendothelial and subepithelial basement membranes (present also in cardiac and skeletal muscles) and especially in vessels with fenestrated endothelium, like those in endocrine glands, liver sinusoidal cells, renal glomeruli, and peritubular capillaries (43), whereas the related collagen XV is restricted to continuous endothelium and its endostatin domain has been termed restin (40). Endostatin is generated from collagen XVIII by the action of proteases, such as cysteine proteinases (cathepsin L), matrix metalloproteinases (−3,−7,−9,−13, −14, and −20), and the serine protease elastase (15–16). Tumor suppressor p53 induces synthesis of collagen XVIII and proteolytic cleavage of the COOH-terminal fragment endostatin (30, 45). Endostatin is a powerful inhibitor of endothelial cell proliferation; it blocks all functions of endothelial and endothelial progenitor cells (migration, survival, vessel stabilization, and angiogenesis) and induces endothelial cell apoptosis. Furthermore, endostatin binds to receptors 1–3 for vascular endothelial growth factor and interferes with its signaling (23). For those reasons we focused attention on endostatin as a plausible mediator of aging-related microvascular dropout and renal fibrosis. The goals of the present study were to 1) screen levels of endostatin in aging mice and 2) correlate them with the progression of microvascular dropout and renal fibrosis. Our findings are consistent with the hypothesis on elevation of endostatin levels in aging and parallel microvascular rarefaction and induction of renal fibrosis.

MATERIALS AND METHODS

Experimental animals.

Studies were conducted in FVB and C57 mice strains, age between 1 to 24 mo old. FVB strain mice were divided into three groups: 1-mo group (n = 5), 10-mo group (n = 2), and 15-mo group (n = 7). C57 mice were collected from two separate sets. There are two age groups including 2- to ∼3 mo group (n = 4–5) and 20-to ∼22 mo group (n = 4–5) in one set. There are 6-mo-old group (n = 6–7) and 24-mo-old group (n = 6–7) in another set. The animal study protocol was in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and was approved by the Institutional Animal Care and Use Committee of New York Medical College.

Serum endostatin measurements.

Serum samples were collected from indicated age groups. Serum Endostatin concentration was measured using mouse endostatin-specific ELISA (USCN Life Science).

Quantitative PCR analysis.

Total RNA was isolated using Spinsmart RNA Mini purification kit (Denville Scientific, Metuchen, NJ). One microgram of RNA was reverse transcribed to single strain cDNA using High Capacity RNA-cDNA kit (Applied Biosystems, Foster City, CA). Real-time quantitative PCR was performed using Perfecta SYBR Green FastMix on a Stratagene MX3000P. Values were normalized for the abundance of the amplified 18S rRNA in each experiment. Fold change in gene expression was determined using the ^2−ΔΔC_T method. The primers (Invitrogen, Grand Island, NY) used are listed in Table 1.

Table 1.

Primer sequences

p16
Forward	AAC TCT TTC GGT CGT ACC CC
Reverse	GCG TGC TTG AGC TGA AGC TA
Endoglin short isoform
Forward	TGA GTA TCC CAA GCC TCC ACC CCA T
Reverse	CTG AGG GGC GTG GGT GAA GGT CAG
Endoglin long isoform
Forward	GCA CTC TGG TAC ATC TAT TCT CAC ACA CGT GG
Reverse	GGG CAC TAC GCC ATG CTG CTG GTG G
SIRT1
Forward	GCT GGA TGA TAT GAC GCT GTG GCA
Reverse	TCC TGC AAC CTG CTC CAA GGT
SIRT3
Forward	TCA CAA CCC CAA GCC CTT TT
Reverse	GTG GGC TTC AAC CAG CTT TG
18S
Forward	AAG GAG ACT CTG GCA TGC TAA C
Reverse	CAG ACA TCT AAG GGC ATC ACA GAC

Western blotting.

Kidney samples were lysed in RIPA buffer. Lysates were clarified by centrifugation, and protein concentration was determined by BCA protein assay (Thermo Scientific, Rockford, IL). Proteins were separated on 4–20% SDS-PAGE gels (Bio-Rad Laboratories, Hercules, CA) and transferred to PVDF membrane using standard condition. Membranes were blocked with 1% BSA in 1× PBS/Tween-20 (washing buffer, pH7.4, with 0.1% of Tween 20) and incubated overnight in the same buffer with primary antibodies added. Primary antibody were used at the dilutions suggested by the manufacturers, which included polyclonal goat anti-endostain (R&D Systems, Minneapolis, MN), rabbit anti-sirtuin 1 and 3 (SIRT1 and SIRT3; Abcam, Cambridge, MA), and monoclonal mouse anti-β-tubulin (Sigma, St. Louis, MO). After primary antibody incubation, membranes were washed and followed by secondary antibodies incubation for 1 h at room temperature. After the excessive secondary antibodies were washed off, the protein bands were visualized using an Immobilon Western chemiluminescence horseradish peroxidase (HRP) substrate (Millipore, Billerica, MA). Secondary antibody used were HRP-conjugated anti-goat (Abcam, Cambridge, MA), anti-rabbit, and anti-mouse (GE Healthcare, Pittsburgh, PA) IgG antibody.

Masson's trichrome staining.

Kidney samples were collected from young and aged animals, fixed in 4% paraformaldehyde, and embedded in paraffin. Paraffin sections (4-μm thick) were stained with Masson's trichrome, and the fibrotic tubulointerstitial areas were quantified using the grid method (266 squares).

CD31 immunohistochemical staining.

Paraffin-embedded kidney sections were used for CD31 staining to detect vasculature endothelium. Following deparaffinization, slides were incubated in antigen retrieval solution (Dako, Carpinteria, CA) at 95–100°C for 30 min and the allowed to cool to room temperature. Slides were blocked with peroxidase, incubated with avidin and biotin blocking solution (Dako), and blocked with 1% BSA for 1 h at room temperature. CD31 antibody (BD Biosciences, San Jose, CA) was applied overnight at 4°C. Slides were stained using Dako LSAB+ system-HRP kit, developed with DAB, and counterstained with hematoxylin. Quantification of CD31 density was performed using grid method (266 squares) applied to images obtained at ×400 magnification.

Laser Doppler imaging of the hindlimb circulation after osmotic pump endostatin peptide delivery.

Endostatin peptide mP1(amino acid sequence: HTHQDFQPVLHLVALNTPLSGGMRGIR), encompassing the 1–27 amino acid portion of mouse endostatin, was synthesized and purified by HPLC (GenScript, Piscataway, NJ). The purity of mP1 peptide was >95%. Peptide was dissolved in 100% DMSO followed by one to one dilution with 1× PBS, and the final concentration of endostatin peptide was 10 mg/ml in 50% DMSO. For the control group, vehicle solution, 50% DMSO was used. It was delivered by osmotic minipump (Alzet, Cupertino, CA) at constant rate of 0.25 μl/h implanted subcutaneously for 28 days. The dosage of peptide was 2.9 mg·kg⁻¹·day⁻¹ to 2-mo-old C57 mice (Jackson Laboratory, Bar Harbor, ME). At day 19 of endostatin peptide delivery, the left femoral artery was ligated. Preligated and postligated blood flow in hindlimb was measured and followed by additional monitoring on days 5 and day 9 after femoral artery ligation (days 24 and 28 after initiation of endostatin peptide delivery) using laser Doppler imaging scanner and flow probes (PIM II; Perimed, Ardmore, PA).

Detection of patent vessels and CD31 immunofluorescence staining.

After 28 days of endostatin peptide delivery, mice were injected with 100 μl of lycopersicon esculentum lectin-Texas red (LEL-TEXAS RED; 1 mg/ml in 10 mm HEPES and 150 mm NaCl pH 7.5; Sigma) via tail vein 5 min to visualize patent vasculature before anesthetization with an intraperitoneal injection of ketamine (10 mg/100g) and xylazine (1 mg/100 g), followed by perfusion with 1× PBS from left ventricular to wash out the not binding lectin. Kidneys were fixed in 4% paraformaldehyde overnight, transferred to PBS containing 30% sucrose overnight, embedded in OCT (Tissue-Tek, Torrance, CA), and cryosectioned (8-μm thick sections). Sections were blocked with PBS-BSA (1%) and stained with rat monoclonal antibody to CD31 (PECAM-1; BD Pharmingen) followed by AlexaFluor FITC-conjugated goat anti-rat (Invitrogen) secondary antibody, according to the manufacturer's recommendation. Nuclei were stained with Hoechst 33342, trihydrochloride, and trihydrate (Invitrogen). The samples were examined using inverted florescent microscopy (Nikon Eclipse TE-2000U). Images were analyzed using ImageJ, and quantification of CD31-positive and lectin/CD31 double-positive peritubular capillary density was performed using grid method (266 squares) applied to images obtained at ×400 magnification.

Statistical analyses.

All values are expressed as means ± SD. Data were analyzed using unpaired t-test. Differences within age groups were considered statistically significant at P < 0.05.

RESULTS

Endostatin levels in young and aging mice.

First, we sought to test the premise that endostatin levels are increased in aging mice. Toward this end, we screened serum samples of young and aged mice. Analysis of serum endostatin levels in C57 mice (2- to ∼3 and 20- to ∼22-mo-old group) using ELISA demonstrated a twofold significant increase in 20- to ∼22-mo-old animals (0.74 ± 0.23 ng/ml) compared with 2- to ∼3-mo-old (0.34 ± 0.09 ng/ml). Whole kidney lysates showed a 2.9-fold elevation in immunodetectable endostatin (Fig. 1A) in 24-mo-old C57 mouse strain. Similar results for immunotecectable endostatin level in the kidney were obtained in C57 mice (2- to ∼3 vs. 20- to ∼22-mo-old groups, data not shown). This increased expression of endostatin in the serum of 20- to ∼22-mo-old mice and in the kidney of 24-mo-old mice was associated with the expression of other markers of senescence: elevation of p16 expression and increased ratio of short:long endoglin (27), as shown in Fig. 1B.

Fig. 1.

Endostatin protein expression in aging kidney. A: whole kidney lysates from young and aged mice showed elevation in immunodetectable endostatin in 24-mo-old C57 mice. B: cell senescence markers including p16 and the ratio of short:long endoglin were increased as evidenced by quantitative PCR in 20- to ∼22-mo-old C57 mice. C: elevation of kidney endostatin protein level in 15-mo-old FVB mice is indicated by immunoblotting. Note that arrow indicating the lower band shows the internal control β-tubulin and the upper band is matrix metallopeptidase 14, a protein not discussed in the current article. Data are means ± SD. *P < 0.05, **P < 0.01; n = 2–7 animals per group.

To confirm these observations in a different mouse strain, parallel experiments were performed in 1-, 10-, and 15-mo-old FVB mice. Longitudinal analysis of endostatin protein expression in the kidneys obtained from FVB mice disclosed an elevation commencing already at 15 mo of age and reaching a 5.6-fold increase (Fig. 1C). Note: the results for 10-mo-old mice are preliminary observations made from two mice throughout Figs. 1–4.

Fig. 4.

Sirtuins 1 and 3 (Sirt1 and Sirt3) protein and gene expression in aging mice kidney. A and B: SIRT1 and SIRT3 mRNA abundance and protein expression were decreased in 20- to 22-mo-old C57 mice as demonstrated by quantitative PCR and Western blotting. C: SIRT1 protein expression in the kidney was decreased in 15-mo-old FVB mice as demonstrated by Western blotting. Data are means ± SD. *P < 0.05, **P < 0.01, ***P < 0.001; n = 2–7 animals per group.

Microvascular density decreases and tubulointerstitial fibrosis increases in aging kidneys.

In view of potent antiangiogenic properties of endostatin, we next inquired whether its upregulation in the aging kidneys is associated with any changes in microvascular density. Immunohistochemical analysis of kidneys revealed that the number of CD-31-positive peritubular capillaries was decreased in kidneys of aged C57 mice (Fig. 2A). Similar observations were made in FVB mice: longitudinal analysis of density of CD-31-positive peritubular capillaries in the kidney disclosed a gradual decline by 15 mo of age (Fig. 2B).

Fig. 2.

Peritubular microvascular density in aging kidneys in C57 mice (A) and FVB mice (B). A, left: images of immunohistochemical analysis revealed that the number of CD-31-positive peritubular capillaries was decreased in the kidneys collected from 20- to 22-mo-old C57 mice and 15-mo-old FVB mice as indicated in B. A, top: representative images at magnification of ×400. Middle: enlargement of rectangle marked area from images at top. Black arrow indicates CD31-positive peritubular capillaries. Right: quantification of CD31 density was performed using a grid method. Data are means ± SD. *P < 0.05, **P < 0.01; n = 2–7 animals per group.

The above upregulation of endostatin expression and microvascular rarefaction in the kidneys of aging mice was accompanied by nephrosclerosis. Masson's trichrome staining of kidneys demonstrated that the area of tubulointerstitial fibrosis was significantly increased in 24-mo-old C57 mice (Fig. 3A). It is remarkable that all these changes in endostatin and senescence markers, as well as development of tubulointerstitital fibrosis, were detectable in a mouse strain, C57, which is least affected by profibrotic stimuli, thus strengthening the initial prediction (14, 25, 50). Tubulointerstitial fibrosis in aged FVB mice was also analyzed and was found significantly elevated already in 15-mo-old FVB mice as indicated by Masson's trichrome staining (Fig. 3B). A direct comprising of the degree of interstitial fibrosis between our FVB and C57 mice group is difficult due to age differences.

Fig. 3.

Interstitial fibrosis in aging mouse kidney C57 mice (A) and FVB mice (B) kidneys show increased interstitial fibrosis in aging mice as analyzed by Masson's trichrome staining. Representative images for each group are shown at magnification of ×100 (top) and ×400 (bottom). Quantification of interstitial fibrosis area was performed using grid method on ×100 magnification images, and results are present at right. Data are means ± SD. *P < 0.001; n = 2–7 animals per group.

SIRT1 and SIRT3: aging-related gene and protein expression.

In the previous studies (47) we have demonstrated that reduction in SIRT1 expression in vascular endothelia results in accelerated fibrosis. Therefore, we next examined expression of SIRT1, as well as SIRT3, at the messenger RNA and protein level. Results of quantitative PCR and Western blotting demonstrated that mRNA abundance for SIRT1 and SIRT3 was reduced in 20- to ∼22-mo-old C57 mice (Fig. 4A) and this was accompanied by the decreased expression of the corresponding proteins (Fig. 4B). Longitudinal analysis of SIRT1 protein expression in the kidney disclosed a gradual decline at 15-mo-old in FVB mice (Fig. 4C). Similar results were obtained in the left ventricular myocardium in FVB mice (data not shown).

Endostatin peptide-treated mice show delayed recovery of circulation after femoral artery ligation.

A previous study showed that the NH₂-terminal domain of endostatin is the functional domain responsible for inhibition of angiogenesis (46). Therefore, we inquired whether the delivery of NH₂-terminal region of endostatin (endostatin peptide mP1 HTHQDFQPVLHLVALNTPLSGGMRGIR, encompassing the 1–27 amino acid portion of the mouse endostatin) could impair angiogenesis and thus mimic the situation of aging kidney. In Fig. 5, A and B, endostatin peptide was delivered by osmotic minipumps implanted subcutaneously for 28 days at the dose of 2.9 mg·kg⁻¹·day⁻¹ to 2-mo-old mice. The dose and time point of endostatin peptide were selected and adjusted according to Skovseth et al. (44) in which 2.3 mg·kg⁻¹·day⁻¹ of recombinant mouse endostain was used for 28 days. After 19 days of endostatin peptide delivery, the left femoral artery was ligated and thereafter laser Doppler flowmetry was performed. The result showed that preligated and postligated blood flow was no significant different between control (50% DMSO) and endostatin peptide-delivered mice. However, the recovery rate of blood flow differed between control and endostatin treated mice. In control mice, recovery of circulation commenced on day 5 and arrived at ∼60% of the initial blood flow. In contrast, recovery of hindlimb circulation was postponed with only 40% blood flow perfusion in endostatin peptide-treated mice on day 5, and hindlimb perfusion remained significantly decreased on day 9, indicating that elevated levels of NH₂-terminal region of endostain interfere with angiogenesis and restoration of collateral circulation under conditions of increased demand.

Fig. 5.

Laser Doppler imaging of the hindlimb circulation in control and endostatin peptide-treated 2-mo-old mice after femoral artery ligation. A: after the femoral ligation (left leg), the circulation recovery was monitored by laser Doppler imaging-scanning. B: quantification of circulation recovery. Preligation served as 100% perfusion baseline reference, and postligation was reflected by reduction of perfusion (percentage). Note that in control mice, circulation recovery commenced on day 5 and remained better than endostatin peptide-treated group on day 9 after femoral artery ligation. Data are means ± SD. *P < 0.05; n = 4–5 animals per group.

Patency of the peritubular microvasculature is impaired in endostatin peptide-treated mice.

After 28 days of endostatin peptide delivery, Lycopersicon esculentum (tomato) lectin was injected via tail vein exactly 5 min before euthanizing animals. Kidneys were removed and analyzed for microvascular density, patency, and degree of fibrosis. Although capillary density did not change significantly in the kidneys of endostatin peptide-treated mice, as evidenced by CD31 immunofluorescence staining, the patency of the vessels was impaired, as indicated by decreased ratio of tomato lectin staining to CD31 staining (Fig. 6, A and B). In addition, a mild patchy tubulointerstitial fibrosis was observed, which did not reach statistical significance (data not shown), and was not associated with functional abnormalities (proteinuria, serum creatinine, data not shown) in endostatin peptide-treated mice.

Fig. 6.

Kidney peritubular capillary density and patency in control and endostatin peptide-treated mice. A: capillary density of the kidney cryosections of the cortex as reflected by the CD31 fluorescent staining. The patency of the peritubular capillaries is indicated by the proportion of the lectin-Texas red-costained capillaries (injected via tail vein 5 min before the animal was killed). Top: images of each treatment group are shown at magnification ×400, and marked rectangle area was enlarged for better visualization and is presented at the bottom of each corresponding image. White arrow: the overlapping area of CD31 and lectin staining. White arrowhead: CD31 positive but lectin negative. B: quantification was performed using grid methods. Data are means ± SD. *P < 0.01, CD31 staining; n = 4 animals per group.

DISCUSSION

Data presented herein show that aging mice of two different strains exhibit highly elevated circulating levels and renal tissue expression of endostatin. In parallel, renal expression of markers of senescence (p16, short:long endoglin) is increased with aging. This is accompanied by the suppressed expression of SIRT1 and SIRT3, decreased microvascular density, and increased tubulointerstitial fibrosis in aging animals.

Guarente and Franklin (18) provided an overview of the roles played by sirtuins in aging. SIRT1 and SIRT3 are major regulators of longevity and mitochondrial metabolism, and both are interlinked: SIRT1 reduces expression of the angiotensin receptor AT₁, lowers the level of reactive oxygen species, and increases longevity by elevating levels of SIRT3 (5, 31). We have recently demonstrated that various cardiovascular risk factors, which lead to endothelial dysfunction, also reduce the expression of SIRT1 (11). Our results obtained in aging mice show decreased gene and protein expression of SIRT1 and 3 (Fig. 4, A and B).

Tissue fibrosis is a common manifestation of aging-associated diseases, including renal fibrosis, which is a subject to individual variations (2, 12, 29, 49). Indeed, in our aging mice from two different strains, tubulointerstitial fibrosis and microvascular rarefaction are conspicuous. It is interesting that this morphologic presentation could not be achieved by a month-long treatment of young animals with endostatin, as shown in Fig. 6. Although no significant tubulointerstitial fibrosis and microvascular rarefaction in the kidney of endostatin-peptide treated mice could be achieved, the patency of the vessels is impaired as indicated by lectin-CD31 costaining (Fig. 6, A and B). In contrast to the kidney, collateral circulation of hindlimb from endostatin-treated mice subjected to femoral artery ligation shows a significant delay in recovery (Fig. 5, A and B). Comparing the results of kidney and hindlimb circulation, it appears that the consequences of elevated endostatin levels are revealed better upon imposing a challenging stimulus rather than under static conditions. It may be also true that the 28-day duration of observations is not sufficient and longer exposure to endostatin study may reveal additional information. The earliest vascular abnormality in endostatin-treated young mice is confined to reduced vessel patency without significant pruning of peritubular capillaries. It is conceivable that reduced microvascular patency is just a prelude to the future microvascular rarefaction. However, the presented set of data, although it shows a correlation between the elevated endostatin levels and increased microvascular rarefaction/interstitial fibrosis in aging mice, does not establish the causality and does not provide definitive proof of a mechanistic link. In the future, we are planning to extend the duration of administration of endostatin peptide using different doses and to use genetically engineered mice over- and underexpressing endostatin to define its role in renal fibrosis.

Increased local production of collagen XVIII and endostatin has been documented in several pathologic conditions, such as liver cirrhosis, ischemic cerebral events, and hypobaric hypoxia (3, 33, 37); however, these conditions are distinct from the normal aging process. We hypothesize that imbalance between pro- and antiangiogenic factors with predominance of the latter may herald dormant tumor(s). There is growing awareness of the clinically silent, nongrowing, or slowly growing tumors in aged mammals. Careful pathologic analysis of >2-yr-old mice revealed high incidence of different tumors like lymphomas, carcinomas, and adenomas in 8–68% of animals (21). Recent studies emphasize that dormant solitary tumor cells or micrometastases may remain quiescent for extended periods of time when their blood supply is limited due to the intrinsic production of angiogenesis inhibitors, like endostatin, thus resulting in “angiogenic dormancy” (34, 38). In the proposed scenario, with advancing age, the accumulating number of dormant transformed cells throughout the body could be responsible for elevated levels of endostatin (perhaps, together with other antiangiogenic factors), which ensures their dormancy. As a matter of fact, another study indicates that aging exhibits impaired angiogenesis and healing because of imbalance between pro- and antiangiogenic factors. It showed a significant increase of endostatin and decreased VEGF protein level in gastric mucosa of aging rat (1). The clinical relevance of both our finding is supported by a recent study showing that increased age is associated with increased serum endostatin concentration (7).

It is, therefore, quite plausible that the same mechanism that underlies tumor dormancy is involved in vascular pruning or insufficient angiogenesis, thus resulting in a widespread fibrosis of aging. It is impossible to prove this statement experimentally at the present time, because such premalignant cells are scattered and their identification would require major detective capabilities. Our data provide a possible scenario that the increase in endostatin with age might be a defense mechanism preventing tumor cell expansion.

The proposed view on the duality of endostatin actions in maintaining tumor dormancy and compromising antifibrotic mechanisms is a modification of the theory of antagonistic pleiotropy (17, 48a), placing emphasize on the role of cellular senescence in tumor suppression and aging-associated pathologies (9–10). In fact, endostatin, which accelerates endothelial cell senescence, could serve as a molecular pivot of antagonistic pleiotropy.

GRANTS

This study was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grants DK-54602, DK-052783, and DK-45462 and Westchester Artificial Kidney Foundation.

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

Author contributions: C.H.S.L., J.C., B.D.Z., J.M., and M.S.G. conception and design of research; C.H.S.L., J.C., B.D.Z., S.M., and J.M. performed experiments; C.H.S.L., J.C., and J.M. analyzed data; C.H.S.L. and J.C. interpreted results of experiments; C.H.S.L. and J.C. prepared figures; C.H.S.L. and M.S.G. drafted manuscript; C.H.S.L., J.C., and M.S.G. edited and revised manuscript; C.H.S.L., J.C., and M.S.G. approved final version of manuscript.