Microvasculature remodeling in the mouse lower gut during inflammaging

Inflammaging is defined as low-grade, chronic, systemic inflammation in aging, in the absence of overt infection. Age-associated deterioration of gastrointestinal function could be ascribed to the inflammaging, although evidence is yet to emerge. Here we show that microvessels in aging mouse intestine were progressively deprived of supportive structures, microvessel-associated pericytes and adherens junction protein vascular endothelial (VE)-cadherin, and became leaky. This alteration was ascribed to up-regulation of angiopoetin-2 in microvascular endothelial cells. Up-regulation of the angiopoietin-2 was by TNF-α, originated from M2-like residential CD206+ macrophages, proportion of which increases as animal ages. It was concluded that antigenic burdens encountered in intestine throughout life create the condition of chronic stage of inflammation, which accumulates M2-like macrophages expressing TNF-α. The TNF-α induces vascular leakage to facilitate recruitment of immune cells into intestine under the chronic inflammatory setting.

Up-regulation of ANG2 in EC by TNF-α secreted by M2-like macrophages in aging animals. As macrophage infiltration has been proposed to be a hallmark of age-associated chronic inflammation 2 , we reasoned that the source of ANG2 inducing factor may be the macrophages in the LP. Quantification of F4/80 + macrophages in LP and also in the entire intestine revealed no age-dependent change (data not shown). The macrophages are broadly categorized into two main subtypes, CD206pro-inflammatory M1 and CD206 + anti-inflammatory M2. M1 macrophages with pro-inflammatory cytokines are important for host defense, but could cause considerable damage to the host. Innate signals such as interleukin 4 (IL-4) are known to differentiate resident macrophages into M2 subtype expressing mannose receptor (CD206) 15 . To determine the distribution of CD206 + and CD206 − cells among F4/80 + macrophages in the LP of entire intestine, we performed fluorescence activated cell sorting (FACS) analysis ( Fig. 2A and B). Fraction of the CD206 + macrophages increased, while CD206 − macrophages concomitantly declined as animal aged: the ratio of CD206 + macrophages vs CD206 − macrophages was about 3:5 in young mice, but later changed to 4:1 at old age. Analysis of gene expression profile of the F4/80 + MHCII + macrophages obtained by FACS revealed increased expression of M2-associated cytokines, CCL2, arginase, and anti-inflammatory IL-10, and decreased expression of M1-associated IL-6 and IL-1β , as animal aged (Fig. 2C) 16 . Notably, TNF-α , a signature of systemic inflammation known to be expressed in M1 subtype, was up-regulated in the F4/80 + MHCII + macrophages of the aged mice 17 . This was further verified by determining TNF-α in intestinal lysate and the macrophages isolated from LP of young and old mice ( Supplementary Fig. 5). Over 2-fold increased expression was detected in both samples from old mice. To identify the source of TNF-α , CD206 + and CD206 − cells among F4/80 + macrophages from young and old mice were separated and analyzed for TNF-α expression. TNF-α expressions from CD206 + cells from old animals was increased about 2-fold compare with that from young animals while that from CD206 − cells were more or less the same (Fig. 2D). These results suggested that gut-resident macrophages in aging animal may not readily fit into the typical classification M1-M2 paradigm. Moreover, expression of M1-associated VEGF-A was reduced in the aged animal as well 16 . Taken together, we hypothesized that TNF-α produced by resident macrophages in aging animals would be important for the up-regulation of ANG2 in ECs, resulting in the deprivation of pericytes that in turn elicits leaky vasculature in LP of aged intestinal villi.
We further corroborate our hypothesis by determining expression of ANG2 from Human Umbilical Vein Endothelial Cells (HUVEC) treated with TNF-α or co-cultured in transwells with macrophages isolated from the intestines of young or old mice (n= 4). TNF-α up-regulated the ANG2 expression by ~3-fold, while co-culturing with macrophages from old mice up-regulated ANG2 by ~2.5-fold compared to that from young mice (Fig. 3A). This up-regulation was nullified by treatment with a specific inhibitor etanercept, a recombinant soluble TNF-α receptor (Fig. 3B) 18 . We then blocked the TNF-α function in old mice using the etanercept. Old animals that  was reinstatement of pericytes. (Fig. 1E shows quantifications). EC integrity was also determined using 70-kDa TRITC-dextran and found to be restored as well (bottom two rows in Fig. 1D). By comparison, no sign of restoration was observed in the aged mice that received IgG, a negative control. It has been reported that TNF-α initiates vascular remodeling by priming EC for sprouting and promoting pericyte recruitment when secreted by activated macrophages 16 . Under this chronic inflammatory setting, however, TNF-α from resident macrophages is playing an adverse role by inducing ANG2 expression in EC, resulting in increased vascular permeability.

Discussion
Homeostatic status of the intestine depends on a complex interplay between age-associated gut microenvironment changes and repopulation of resident immune cells 19 . It has been reported that the total number of F4/80 + CD11b + macrophages in the mouse colon is little at birth, but increases progressively to adulthood (5 to 9 weeks) in microbiota-dependent manner 20 . In this study, we propose that the antigenic burden encountered in the intestine causes macrophage infiltration during the first few months after birth and that is sustained throughout life (Fig. 4). Under the condition of chronic inflammation, it stands to reason to polarize macrophage toward the M2-like subtype to avoid tissue injury and eventual chaotic consequences caused by activated M1 macrophage. Consistently, it has been reported that total macrophages and myeloid derived suppressor cells cumulated in the spleens and bone marrow of aged mice were mostly anti-inflammatory M2 cells in aged mice 21 . Macrophages are the sources of both pro-angiogenic and anti-angiogenic factors, which can differentially guide vascular network formation under many pathological conditions. We therefore propose that TNF-α derived from the macrophages in aged animals skews the angiopoetin-TIE-2 signaling in vascular EC to inflammatory settings that would facilitate recruitment of immune cells through ECs. Such increase in vasculature permeability entails modulation of EC network such as loss of VE-cadherin and pericyte 22 , as demonstrated in this study. Together, our study demonstrates for the first time, to the best of our knowledge, that sustained aggravation of inflammation leads to age-related structural changes in organ.

Materials and Methods
Mice. Female mice at 3-month, 12-month and 22-month (or older) old aged C57BL/6 J were used for this study. All mice were purchased from Animal Facility of Aging Science in Korea Basic Science Institute (Gwangju, Korea). Animals were anesthetized by intramuscular injection of a combination of anesthetics (80 mg/kg ketamine and 12 mg/kg xylazine) and sacrificed. All animal experiments were approved and performed following the Q-PCR analysis. Total RNA was isolated from flow cytometry sorted cells or small intestine using the TRIzol reagent (Invitrogen), according to the manufacturer's instructions. mRNAs were reverse transcribed to cDNAs with a RT PreMix Kit (Enzynomics) and then analyzed by the real-time PCR system (Qiagen) using SYBR Green PCR Master Mix (Enzynomics). All data were normalized to GAPDH expression. Primer sets used are shown in Supplementary Table 1.

Preparation of resident macrophages in LP.
Resident macrophages in LP were isolated by collagenase dissociation method as described previously 9 . The entire intestine of the mice at different ages was cut into 3-4 pieces and then inverted on polyethylene tubes (Becton Dickinson), washed three times with calcium-and magnesium-free phosphate buffered saline (PBS; Lonza) and mucus was removed with 1 mM dithiothreitol (DTT; Sigma-Aldrich). The intestinal epithelium was eluted with 30 mM EDTA at room temperature, followed by digestion with 36 U/ml type IV collagenase (Sigma-Aldrich) in Dulbecco's Modified Eagle's Medium (DMEM; Lonza) containing 5% fetal bovine serum (FBS) for 90 minutes at 37 °C in a 5% CO 2 humidified incubator. The digested tissue was gently shaken for 10 minutes at room temperature and then passed through a 70 μ m nylon cell strainer and washed with DMEM.
To purify CD11b and F4/80 expressing macrophages, isolated cells were incubated in 10% serum and Fc block,

Vascular leakage assay.
To visualize and determine the vascular leakage, 70-kDa TRITC-dextran (Invitrogen) was used as tracer 14 . The animal was intravenously injected with TRITC-dextrans (2 mg/ 20 g of body weight). Thirty min after injection, the mice were anesthetized and transcardial perfusion was performed with 10 ml of saline to remove remaining tracer from vessels. The small intestine was collected and fixed with 10% neutralized formalin. Confocal microscope image was acquired after stain of vascular network with anti-CD31 antibody.
To quantify the vascular leakage in small intestine, mice were treated as described above and small intestine was homogenized using liquid nitrogen, resuspended in TBS containing 0.1% Triton X-100. After centrifugation, supernatant was analyzed using Fluorometer (TECAN) with excitation/emission wavelengths of 544/590 nm.

Assessment of ANG2 expression levels in HUVEC.
HUVEC grown in EBM2 media were treated with TNF-α (Sigma Aldrich, 10 ng/ml) for 10 hours. Alternatively, HUVEC grown in EBM2 media were co-cultured with macrophages isolated from mouse gut at young and old ages in trans-well plates (Corning, 3 nm/pore) for indicated time. ANG2 expression in the HUVEC was determined by Q-PCR.
Statistical analysis. Data were analyzed using Prism statistical software. The data met assumptions of a normal distribution as determined by statistical software, and variance was estimated with s.e.m. The two-tailed Student's t-test was used for statistical analysis between two groups. Differences were considered statistically significant at p < 0.05.