Chlamydia pneumoniae Lung Infection in Mice Induces Fatty Acid–Binding Protein 4–Dependent White Adipose Tissue Pathology

This article is featured in Top Reads, p. 1013

Many infections can alter metabolism, and their pathogenesis may be modified by the metabolic state of the host. For example, the impact of infection on insulin action and glucose homeostasis of the host has been shown in many studies as early as the early 1900s (1, 2). The association of Chlamydia pneumoniae infection with metabolic syndrome has also been studied and documented (3–6). However, whether C. pneumoniae has a causal role in metabolic syndrome and the underlying mechanisms of this interaction remain to be determined (7–10).

White adipose tissues (WATs) are crucial in maintaining whole-body energy homeostasis. Their dysfunction can contribute to hepatic insulin resistance and type 2 diabetes mellitus (11). It has recently been well appreciated that chronic low-grade inflammation, often referred to as meta-inflammation, develops in obese individuals and is crucial for insulin sensitivity (12–14). However, it is still largely unknown whether and, if so, how infectious pathogens induce WAT inflammation and cause systemic metabolic disturbances.

Fatty acid–binding protein 4 (FABP4) is present in adipocytes and macrophages. FABP4 controls essential immunometabolic functions, contributes to adipose tissue inflammation and dysfunction, and links obesity and obesity-related disorders, such as atherosclerosis and insulin resistance (15–17). Interestingly, a new hormone called fabkin, where secreted FABP4 is the core component, has been demonstrated in recent studies to establish direct endocrine links between adipose tissue and distant organs (18). Genetic, chemical, or pharmacological targeting of FABP4 results in marked protection against cardiometabolic disease (19, 20). Multiple independent genome-wide association studies in humans carrying a haploinsufficiency allele of FABP4 show that these individuals have the same phenotype (21).

We have previously reported that C. pneumoniae, which needs to obtain nutrients, such as ATP and lipids from host cells, successfully infects and proliferates in adipocytes by inducing hormone-sensitive lipase (HSL)–mediated lipolysis. Interestingly, C. pneumoniae exploits host FABP4 to facilitate fat mobilization and intracellular growth in adipocytes, implying that intracellular pathogens acquire energy by hijacking the lipid metabolism pathways of their hosts (22). We also reported that C. pneumoniae infection–induced endoplasmic reticulum (ER) stress leads to the robust secretion of FABP4 from adipocytes in vitro. Such a process may provide new insights into the etiological link between C. pneumoniae infection and metabolic syndrome (23). However, the relevance of these findings to an in vivo context is unclear. Thus, it remains to be determined whether C. pneumoniae in vivo lung infection induces histopathology in fat tissues and causes metabolic disturbances.

We demonstrate that C. pneumoniae resides inside adipocytes and adipose tissue macrophages (ATMs), inducing vigorous lipolysis and robust inflammation in WATs after intranasal infection with C. pneumoniae in mice. We demonstrate that FABP4 deficiency abrogates C. pneumoniae infection–induced WAT pathology. Furthermore, in vivo treatment with azoramide, a modulator of the unfolded protein response (UPR), decreased C. pneumoniae infection–induced WAT pathology. Our results provide critical new insight into the underlying mechanism of metabolic pathologies induced by infectious agents, such as C. pneumoniae.

C57BL/6J wild-type (WT) mice were purchased from Japan SLC. FABP4^−/− mice (24) were backcrossed for 12 generations to C57BL/6J mice (20) and C57BL/6J-background FABP4^−/− mice were housed in a specific pathogen-free facility. All animal experiments were conducted following protocols approved by the Animal Care and Use Committee of Fukuoka University.

C. pneumoniae (strain AR39, ATCC 53592) was obtained from the American Type Culture Collection and propagated as previously described (25). Chlamydial elementary bodies (EBs) were purified using Urografin (Bayer) density gradient centrifugation, resuspended in sucrose-phosphate-glutamate (SPG) buffer, and stored at −80°C. All Chlamydia stocks were confirmed negative for Mycoplasma contamination using a MycoAlert Mycoplasma detection kit (Lonza).

Reagents were obtained from the following sources: BMS309403 was provided by Bristol Myers Squibb or obtained from AdipoGen Life Sciences. The following Abs were used: monoclonal mouse anti–β-actin (Santa Cruz Biotechnology, SC-47778), polyclonal rabbit anti–phospho-HSL (Ser⁶⁶⁰) (Cell Signaling Technology, 4126), polyclonal rabbit anti-HSL (Cell Signaling Technology, 4107), monoclonal rabbit anti–inositol-requiring enzyme 1α (IRE1α) (Cell Signaling Technology, 3294), monoclonal mouse anti–C/EBP homologous protein (CHOP) (Cell Signaling Technology, 2895), monoclonal rabbit anti-BiP (Cell Signaling Technology, 3177), monoclonal rabbit anti–phospho-eIF2α (Ser⁵¹) (Cell Signaling Technology, 9721), anti-Chlamydia rabbit Ab (Abcam, ab31131), anti-CD11c (OriGene, TA323539), anti-mannose receptor (Abcam, ab64693), anti-F4/80 (Serotec, MCA497GA), anti–TNF-α (Abcam, ab6671), anti–IL-6 (Abcam, ab6672), Histofine Simple Stain MAX PO (peroxidase) (R) for rabbit primary Ab (Nichirei Biosciences), EnVision+ kit/HRP secondary Ab (Dako), and Oil Red O (Abcam, ab146295).

C57BL/6J WT mice or C57BL/6J-background FABP4 ^−/− mice (female, aged 8–9 wk) were mildly sedated with isoflurane and inoculated intranasally with 2 × 10⁵ to 1 × 10⁶ inclusion-forming units (IFU) of C. pneumoniae strain AR39 in 20 μl of PBS. The mice were euthanized on the indicated days postinfection. The mice were perfused with 20 ml of EDTA (3.0 mg) containing PBS, and their lungs, liver, and gonadal adipose tissues were aseptically isolated. These tissues were homogenized in SPG buffer using a cell grinder. Then, the tissue homogenates were centrifuged at 1900 × g for 30 min at 4°C to remove coarse tissues and debris. The supernatants were removed and stored frozen at −80°C. The tissue homogenates were serially diluted with SPG buffer and added in duplicate to a monolayer of HeLa cells to quantify infectious C. pneumoniae in tissues. Then, the IFU were determined and expressed as the number of IFU per organ.

The stromal vascular fraction (SVF) cells and adipocytes were isolated from the WAT of C. pneumoniae–infected mice. Briefly, gonadal and omentum fat pads were minced in PBS and digested with 2 mg/ml collagenase type II (C6885; Sigma-Aldrich) at 37°C for 1 h. Then, the digested tissue was filtered through a nylon mesh and centrifuged at 500 × g for 10 min. The top layer (adipocyte fraction) was collected and mixed with DMEM/10% FCS. The remaining SVF pellet was treated with ammonium-chloride-potassium lysis buffer at 37°C for 5 min to lyse RBCs and resuspended in DMEM/10% FCS. Next, the cell suspension was filtered through a Falcon cell strainer (40 μm, BD Falcon), centrifuged at 500 × g for 10 min, and resuspended in DMEM/10% FCS. After counting the number of cells, both the SVF and adipocyte suspensions were mixed with SPG, sonicated, centrifuged at 12,000 × g at 4°C for 30 min, and resuspended in SPG. The SVF and adipocyte suspensions were serially diluted with SPG and added to a monolayer of HeLa cells to quantify infectious C. pneumoniae EBs. Then, the number of IFU was determined and expressed as the number of IFU per tissue.

All Chlamydia-infected cells were collected, frozen, and thawed, serially diluted 10-fold in SPG medium, and reseeded into 24-well plates containing a monolayer of HeLa cells (2 × 10⁵ cells per well). After 24–48 h, the cells were fixed with ice-cold methanol and incubated with a FITC-conjugated anti-Chlamydia LPS-specific mAb (PROGEN) and DAPI. The growing chlamydial inclusions were counted using Axioskop fluorescent microscopy (Zeiss). Then, the mean number of IFU was calculated.

In ex vivo infection experiments, gonadal WATs were cut with a tissue chopper (McIlwain) into 0.5-mm squares and cultured on transwell dishes (Kurabo Industries) in DMEM with 10% FBS and insulin-transferrin-sodium selenite liquid media (Sigma-Aldrich, 13146) (26). After overnight culture, WATs were infected with the indicated volume of 3.3 × 10⁷ IFU/ml C. pneumoniae for 3 d and used for further analyses.

The 3T3-L1 adipocytes were differentiated as previously described (22, 23). The medium was then replaced by DMEM supplemented with 10 μg/ml insulin for 8 d. Matured 3T3-L1 adipocytes were seeded into 24 well-dishes and used for further analysis.

The mouse FABP4 gene was cloned into the His-tag expression vector pET-15b. The vector was transformed into BL21 DE3 Escherichia coli. FABP4 protein expressions were induced with isopropyl-β-d-thiogalactopyranoside and purified with a Ni-NTA column (Qiagen). Endotoxin was removed with a high-capacity endotoxin removal spin column (Pierce) and confirmed with a ToxinSensor chromogenic LAL (Limulus amebocyte lysate) endotoxin assay kit (GenScript, value <0.1 endotoxin unit/μg protein). Purified recombinant FABP4 proteins were added to matured 3T3L1 adipocytes at the indicated concentrations.

The transmission electron microscope analysis was outsourced to the Tokai electron microscope analysis. Briefly, isolated WAT was fixed with 2% paraformaldehyde and 2% glutaraldehyde in 0.1 M cacodylate buffer overnight at 4°C. After three washes with 0.1 M cacodylate buffer for 30 min at 4°C, samples were fixed with 2% osmium tetroxide in 0.1 M cacodylate buffer for 6 h at 4°C. Fixed samples were dehydrated with graduated ethanol, infiltrated with resin, and then embedded with Quetol-812 (Nissin EM) for 48 h at 60°C. Thin sections (80 nm) were stained with 2% uranyl acetate at room temperature for 15 min and stained with lead stain solution for 3 min at room temperature. Images were obtained using JEM-1400Plus transmission electron microscopy.

In FABP4 inhibitor experiments, the mice were treated by oral gavage before and during C. pneumoniae infection (from 2 d before infection to day 6 postinfection [p.i.]) with a vehicle including 10% 1-methyl-2-pyrrolidone (Sigma-Aldrich, M79603) and 5% Cremophor EL (Calbiochem) with ethanol in 100 μl of water (vehicle control group) or 15 mg/kg/d FABP4 inhibitor BMS309403 dissolved in the vehicle for 9 d before and during C. pneumoniae infection (19).

Female C57BL/6J WT mice were treated with azoramide (MedChemExpress) once a day by oral gavage (150 mg/kg in 200 μl of vehicle solution) from 1 d before C. pneumoniae infection. The vehicle solution was 10% (v/v) ethanol, 0.1% (v/v) acetic acid, 40% (v/v) polyethylene glycol 400, and 0.05% (w/v) O-carboxymethylcellulose (27). 4-Phenyl butyric acid (4-PBA; LKT Laboratories, St. Paul, MN) was administered orally by gavage at a dose of 10 mg in 100 μl of distilled water every 12 h (28).

The levels of plasma glucose in mice were determined using Glutest Neo Sensor (Sanwa Kagaku Kenkyusho, Aichi, Japan). For an oral glucose tolerance test (GTT), normal chow diet C57BL/6 WT mice or FABP4-deficient mice (8–9 wk of age, females) were intranasally infected with C. pneumoniae and at 6 d p.i. mice were fasted for 16 h and treated by oral gavage with 0.15 g/ml glucose solution (10 μl/g body weight). For insulin tolerance tests, mice were fasted for 5 h and i.p. injected with 0.75 mU/g body weight of Humulin R U-100 (Eli Lilly Japan K.K.). The levels of plasm insulin were measured using ultra-sensitive mouse insulin ELISA kit (Cat. M1104, Morinaga Institute of Biological Science). The homeostatic model assessment of insulin resistance (HOMA-IR) in mice was calculated as follows: (fasting blood glucose [mg/dl] × fasting blood insulin [μU/ml])/405.

We isolated WATs from the mice at day 7 p.i. Then, the tissue was fixed in 4% paraformaldehyde, embedded in a paraffin block, and sectioned. The tissue sections were deparaffinized and dehydrated. Following Ag retrieval and quenching of endogenous peroxidases, the tissue sections were blocked with 10% goat serum (Cedarlane) in PBS for 1 h at room temperature. C. pneumoniae Ags were detected using rabbit polyclonal anti-Chlamydia Ab and EnVision+ kit/HRP secondary Ab. Anti-F4/80 rat Ab, anti-CD11c rabbit Ab, anti-CD206 rabbit Ab, anti–TNF-α rabbit Ab, and anti–IL-6 rabbit Ab were also used in this study. Histofine Simple Stain MAX PO (peroxidase) (R) was used to stain rat polyclonal Abs. All sections were developed with diaminobenzidine and washed. The nuclei were counterstained with hematoxylin. Negative controls without primary Abs were treated as described above. Images were obtained using a BZ-9000 digital charge-coupled device microscope (Keyence). Analyses of the number and size of adipocytes in adipose tissues were conducted using Adiposoft software (ImageJ).

The amount of C. pneumoniae present in the tissues of infected mice was assessed using a previously established quantitative PCR (qPCR) method (29, 30). Briefly, C. pneumoniae genomic DNA was isolated from the lung, liver, and gonadal adipose tissues using TRIzol reagent. The chlamydial 16S DNA-isolated DNA template was then quantified by qPCR on an ABI Prism 7500 real-time PCR system (Applied Biosystems) using probe qPCR mix (Takara Bio). The thermal cycling parameters used were 30 s at 95°C, followed by 40 cycles of denaturation at 95°C for 5 s, and annealing and extension at 60°C for 34 s. The qPCR primer sequences were 16S-qPCR1 (5′-CGACTACATGAAGTCGGAATTG-3′), 16S-qPCR2 (5′-ACCTTGGGCGCCTCTCT-3′), and a dual-labeled 16S probe (5′-[6-FAM]-GGCCTTGTACACACCGCCCGT-[TAMRA]-3′). The amount of chlamydial DNA was calculated using the standard curve from serial dilutions of a known amount of purified C. pneumoniae EBs.

Cells were washed once in ice-cold PBS and lysed in RIPA buffer (50 mM Tris [pH 7.4], 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, 150 mM NaCl, 2 mM EDTA, and 50 mM NaF) with a protease inhibitor mixture (Nacalai Tesque) and PhosSTOP phosphatase inhibitor mixture (Roche). The cells were then briefly sonicated. Protein concentration was determined with a BCA (bicinchoninic acid) protein assay kit (Thermo Fisher Scientific), and equal amounts of proteins were loaded onto SDS polyacrylamide gels. Proteins were transferred to membranes using a Trans-Blot SD semi-dry transfer cell (Bio-Rad), following the manufacturer’s instructions. Membranes were blocked in Blocking One (Nacalai Tesque) for 30 min. Membranes were incubated overnight with primary Ab diluted in Can Get Signal solution 1 (Toyobo) at 4°C. The membranes were then incubated with HRP-conjugated secondary Abs to rabbit/mouse IgG (GE Healthcare) or goat IgG (SeraCare) diluted in Can Get Signal solution 2 (Toyobo) for 1 h at room temperature. Immunoreactive bands were detected with ECL blotting reagents (GE Healthcare, RPN2109) or EzWestLumi plus (ATTO). The densitometric analysis was conducted using Image Studio Lite software (LI-COR Biosciences). The densitometric ratio of each band to total HSL or β-actin was shown below the band of each lane.

Gonadal adipose tissues were isolated from C. pneumoniae–infected or noninfected WT or FABP4^−/− female mice. Total RNA was isolated using Isogen II (Nippon Gene). cDNA synthesis was conducted using a PrimeScript RT reagent kit with genomic DNA Eraser (perfect real time) (Takara Bio) and 1 μg of total RNA as a template. Relative gene expression levels were determined in 96-well plates using the SYBR Premix Dimer Eraser (perfect real time) (TaKaRa Bio) with an Applied Biosystems 7500 real-time PCR system (Applied Biosystems). All primer sequences used in the current study are listed in Table I. The specific thermal cycling parameters were as follows: 30 s at 95°C, 40 cycles of denaturation at 95°C for 5 s, annealing at 55°C for 30 s, and extension at 72°C for 34 s.

The plasma levels of FABP4 were measured on the indicated days after C. pneumoniae infection using a CircuLex mouse FABP4/A-FABP ELISA kit (MBL International). Plasma and gonadal adipose tissue lysates were measured for mouse IL-6 or TNF-α with the DuoSet ELISA development system (R&D Systems). Glycerol levels in the supernatant of 3T3L1 adipocytes, WAT ex vivo cultures, or plasma were measured with a glycerol cell-based assay kit (Cayman Chemical) according to the manufacturer’s instructions.

C57BL/6J WT or FABP4^−/− recipient female mice aged 5–7 wk were irradiated with 9.50 Gy from a cesium source and 2 h later reconstituted with 2 × 10⁶ total (CD3⁺ cells depleted) bone marrow (BM) cells from C57BL/6J or FABP4^−/− female mice. Mice were housed for 8 wk to permit the engraftment of BM cells. Four groups of chimeric mice (WT donor into WT recipient, FABP4^−/− donor into FABP4^−/− recipient, WT donor into FABP4^−/− recipient, and FABP4^−/− donor into WT recipient) were infected with 2 × 10⁵ IFU of C. pneumoniae intranasally (n = 6 mice per group).

Hematopoietic cells from the peripheral blood of chimeric mice were genotyped to ensure the successful reconstitution of donor BM. DNA was isolated from blood and tails. PCR was performed to determine the genotype of FABP4 using three specific primers (5′-CAG CAC TCA CCC ACT TCT TTC AT-3′, 5′-ACA TAC AGG GTC TGG TCA TG-3′, 5′-ATA GCA GCC AGT CCC TTC CCG CCT-3′) (31).

The results are expressed as the mean ± SD. Statistical significance of the differences between the two groups was measured with an unpaired t test with Welch’s correction. For analysis of multiple groups, one-way or two-way ANOVA was used with Tukey’s or Sidak’s multiple comparison tests. Data and linear regression analyses were performed using GraphPad Prism software, version 8.4.2. Statistically significant values are noted as follows: *p < 0.05, **p < 0.01, and ***p < 0.001.

We previously reported that C. pneumoniae infects murine 3T3-L1 differentiated adipocytes and induces lipolysis and FABP4 secretion. We determined whether acute C. pneumoniae lung infection induces lipolysis in WAT in mice to validate our previous in vitro findings. C57BL/6 WT mice were intranasally infected with 2 × 10⁵ IFU of C. pneumoniae. Notably, a significant reduction in WAT weight was observed after C. pneumoniae infection (Fig. 1A). It was revealed with histologic examinations of the WAT of infected mice that adipocytes were markedly reduced in size at day 7 p.i. It was suggested by this result that lipolysis in WAT may have been induced by C. pneumoniae infection (Fig. 1B). We also detected chlamydial inclusion bodies in WATs by immunohistochemical analyses in these infected animals (Fig. 1C). However, the replication of C. pneumoniae in WAT was relatively limited compared with those obtained from lungs by a qPCR assay (Fig. 1D). Notably, FABP4 in plasma increased postinfection with C. pneumoniae on days 5 and 7 of infection (Fig. 1E). These results suggest that C. pneumoniae lung infection targets adipose tissue and causes FABP4 secretion in vivo, which strengthens our previous in vitro findings.

It has been shown that lean FABP4^−/− mice have reduced lipolytic capacity at submaximal stimulation (20). We have also previously reported that C. pneumoniae infection–induced lipolysis in adipocytes was abrogated with chemical inhibition or genetic manipulation of FABP4 (22). Thus, we investigated whether C. pneumoniae lung infection–induced WAT lipolysis is FABP4-dependent. FABP4^−/− mice or control WT mice were intranasally infected with 2 × 10⁵ IFU of C. pneumoniae. Although a slight reduction in WAT weight was also observed in FABP4^−/− mice after C. pneumoniae infection (Fig. 2A), it was revealed with histologic examination of the WAT of infected mice that the infection-induced decrease in adipocyte cell size observed in WT mice was primarily abrogated in FABP4^−/− mice (Fig. 2B). There was not much difference in the appearance of chlamydial inclusion bodies in infected cells of the lung section of WT or FABP4^−/− mice on day 5 p.i. (Supplemental Fig. 1A). Although bacterial burdens in the lung of WT mice on day 5 p.i. were a little higher than those of FABP4^−/− mice, there was no statistically significant difference between WT and FABP4^−/− mice. Of note, a nearly equal degree of bacterial dissemination to the liver from the lung was observed in WT and FABP4^−/− mice after intranasal infection of C. pneumoniae (Supplemental Fig. 1B), implying that similar levels of bacterial dissemination to WAT might also have occurred. Therefore, the difference in the changes in adipocyte cell size between WT mice and FABP4^−/− mice on day 7 p.i. was unlikely due to the different levels of bacterial dissemination into WAT from the lung, indicating the possibility of FABP4-dependent WAT lipolysis in infected WT mice.

Infected WT mice showed a tendency to have lesser food intake, although statistically not significant, and severe body weight loss on days 4 and 5 p.i. reached a significant difference between WT mice and FABP4^−/− mice (Supplemental Fig. 1C, 1D), suggesting that starvation-induced lipolysis might be involved in the different adipocyte cell sizes changes in WT and FABP4^−/− mice after C. pneumoniae infection. However, starvation-induced lipolysis alone did not explain a massive accumulation of inflammatory cells in the WAT of infected WT mice (Fig. 1C). Thus, we prefer to consider that C. pneumoniae lung infection might induce more robust infection-induced and/or inflammation-induced lipolysis in WAT in the presence of FABP4. To further confirm this FABP4-dependent WAT lipolysis in C. pneumoniae–infected mice, we next examined the effect of FABP4 inhibitor on C. pneumoniae infection–induced WAT lipolysis in WT mice.

We found that chemical inhibition of FABP4 by BMS309403 but not vehicle treatment almost completely inhibited WAT lipolysis (Supplemental Fig. 2A). We determined the protein abundance and activation status of HSL, an enzyme central to lipolysis, in gonadal adipose tissues obtained from uninfected mice and C. pneumoniae–infected mice (day 5 p.i.), confirming that C. pneumoniae lung infection induced WAT lipolysis. Expression levels of phosphorylated HSL (pHSL) and pHSL/HSL ratio increased in the WAT of WT mice after C. pneumoniae infection on day 5 p.i. compared with uninfected mice (Fig. 2C). It is noteworthy that HSL activation, judged by the pHSL/HSL ratio, decreased in FABP4^−/− mice compared with WT mice (Fig. 2C, right panel).

In addition to adipocytes, WAT contains numerous other cell types, such as macrophages and T lymphocytes, which constitute the SVF. Thus, we examined whether adipocytes and/or SVF cells in the WAT of C. pneumoniae–infected WT or FABP4^−/− mice harbor chlamydial infectious progeny by IFU assays. As shown in Fig. 2D, we found that the absence of FABP4 caused a significant decrease in chlamydial progeny numbers in WAT, suggesting that FABP4 is required for the optimal growth of C. pneumoniae in WAT. These results indicate that FABP4 deficiency protects against C. pneumoniae–induced lipolysis and decreases the chlamydial progeny numbers in the WAT of mice.

Several reports have shown that ER stress is one of the triggers of lipolytic events in adipose cells (32–34). We have previously reported that C. pneumoniae infection induces ER stress and UPR in 3T3L-1 adipocytes, which leads to lipolysis and FABP4 secretion (23). Thus, we examined whether C. pneumoniae infection could induce lipolysis through ER stress in WAT using ex vivo WAT cultures and infection with C. pneumoniae. In this ex vivo WAT infection model (26, 35), we could successfully detect the expression of chlamydial genes (16S rRNA, groEL), which were mainly expressed in the chlamydial replicating form (reticular body [RB]) (Supplemental Fig. 3A). Furthermore, electronic microscopy revealed the presence of RBs in adipocytes (Supplemental Fig. 3B). These results validated this ex vivo WAT infection model. Immunoblot assays were conducted to clarify the association between infection-induced lipolysis and ER stress/UPR. Increased expression levels of IRE1α and CHOP, which were novel ER stress marker proteins, depending on bacterial dose, were observed (Fig. 3A). The ratio of pHSL to the total HSL was also increased, which was in line with our previous reports (22). Notably, in adipose tissue explant cultures from FABP4^−/− mice, HSL activation and the expression levels of ER stress/UPR markers upon C. pneumoniae infection were abrogated (Fig. 3B). Such a result is an indicator that C. pneumoniae infection–induced ER stress and subsequent lipolysis occurs in a FABP4-dependent manner. Exogenous FABP4 has been shown to promote lipolysis and inflammation in adipocytes (36). We also observed that exogenous FABP4 treatment in differentiated 3T3L1 adipocytes induced ER stress (Supplemental Fig. 3C), followed by lipolysis (Fig. 3C).

Notably, glycerol release from 3T3-L1 adipocytes after treatment with exogenous FABP4 was canceled by the simultaneous treatment with the FABP4 inhibitor BMS309403 (Fig. 3C). These results raise the possibility that FABP4 or fabkin (18) secreted from C. pneumoniae–infected adipocytes may be taken up by other neighboring intact adipocytes or ATMs, which may induce ER stress activation and trigger lipolysis and inflammation, followed by FABP4 secretion, leading to WAT pathology and dysfunction, ultimately resulting in the disruption of metabolic homeostasis.

Some small molecules, such as azoramide (27) and 4-PBA (28, 37), have been identified as modulators of ER stress and chemical chaperon, respectively, which have the potential to prevent several metabolic diseases, such as diabetes or steatohepatitis. We examined the effects of azoramide in mice infected with C. pneumoniae in this study. We found that in vivo treatment with azoramide could partially cancel the reduction of adipocyte cell size, adipose tissue weight, and body weight of mice (Fig. 3D), strongly suggesting that treatment with azoramide partially reverses the phenotype in mice infected with C. pneumoniae. These ex vivo and in vivo results suggest that C. pneumoniae infection–induced WAT lipolysis is induced by ER stress and UPR, depending on FABP4.

We prepared BM chimeric mice (WT→WT, WT→FABP4^−/−, FABP4^−/−→WT, or FABP4^−/−→FABP4^−/− mice) (31). We examined the extent of chimerism by genotyping for the WT and FABP4-deficient alleles using genomic DNA extracted from blood and stromal tissue (tail) and confirmed >94.2% reconstitution of BM-derived blood cells in all of the groups (Fig. 4A). As shown in Fig. 4B (upper panels), FABP4 protein expression was detected both in adipocytes and BM-derived ATMs in WT→WT BM chimeric mice. FABP4 expression in WAT of WT→FABP4^−/− BM chimeric mice was only detected in ATMs but not in adipocytes, whereas FABP4 expression in WAT of FABP4^−/−→WT mice was limited to adipocytes but not to ATMs. We infected these mice with C. pneumoniae to determine which source of FABP4 is required for C. pneumoniae infection–induced WAT pathology (lipolysis and inflammation). As shown in Fig. 4B and 4C, the presence of FABP4 in adipocytes or BM-derived ATMs is sufficient to induce WATlipolysis and inflammation after C. pneumoniae infection. The reduction rate of adipocyte cell sizes upon C. pneumoniae infection was significantly diminished only in FABP4^−/−→FABP4^−/− mice, strongly indicating the importance of FABP4 (Fig. 4C, right panel).

Remarkably, we observed increased expression of FABP4 in WAT after C. pneumoniae infection. Abundant expression of FABP4 in adipocytes and ATMs was observed in C. pneumoniae–infected WT→WT BM chimeric mice (Fig. 4B). Increased expression of FABP4 in the WAT of infected WT mice was confirmed by immunoblot and ELISA (Fig. 4D, 4E). FABP4 was expressed massively in ATMs and less on adipocytes in WT→FABP4^−/− mice after C. pneumoniae infection. In FABP4^−/−→WT mice, FABP4 was expressed on adipocytes and ATMs, implying that secreted FABP4 was taken up by ATMs. These findings support the hypothesis that FABP4 secreted from C. pneumoniae–infected adipocytes may be taken up by neighboring intact adipocytes or ATMs.

In addition to robust lipolysis in WAT, C. pneumoniae lung infection in WT mice induced a massive accumulation of inflammatory cells at this site (Fig. 1C). Thus, we examined whether inflammatory cytokines such as TNF-α and IL-6 are expressed in these inflamed WATs of C. pneumoniae–infected animals. We observed a strikingly increased number of cells expressing TNF-α or IL-6 in the WAT of infected WT mice (Fig. 5A, 5B). Importantly, these increases in TNF-α– or IL-6–expressing cells of WAT were not evident in the WAT of FABP4^−/− mice following infection. In vivo treatment with FABP4 inhibitor (BMS309403) also significantly decreased the levels of TNF-α and IL-6 in WAT lysates of infected mice (Supplemental Fig. 2B). In addition to local inflammatory changes, we also found that plasma TNF-α and IL-6 levels on day 7 p.i. were significantly higher in WT mice than in FABP4^−/− mice (Fig. 5C). Thus, these results strongly suggest that the genetic ablation or chemical inhibition of FABP4 protects against C. pneumoniae infection–induced lipolysis and inflammatory cytokine production in the WAT of mice.

We examined the activation markers of M1 and M2 macrophages by immunohistochemistry in adipose tissues to obtain insight into the nature of inflammation in WAT after C. pneumoniae infection. It has been shown that the M2 subtype is the predominant cell type of ATMs in healthy lean mice, whereas M1 macrophages dominate adipose tissues in obese mice and promote inflammation (38). We found increased numbers of CD11c⁺ M1-like ATMs and decreased numbers of mannose receptor CD206⁺ M2-like ATMs in the WAT after C. pneumoniae infection in WT mice (Fig. 6A). In sharp contrast to this, infection-induced increases in the number of CD11c⁺ ATMs or decreases in the number of CD206⁺ M2-like ATMs were not observed in FABP4^−/− mice, and the predominance of CD206⁺ M2-like ATMs were preserved, even postinfection with C. pneumoniae. These findings were partially confirmed by mRNA RT-PCR analysis showing an increase of TNF-α mRNA in the WAT of WT mice postinfection (Fig. 6B, Table I). Interestingly, mRNA for Arg1 and MMR (macrophage mannose receptor) significantly increased in the WAT of FABP4^−/− mice after C. pneumoniae infection. In line with this finding, mRNA for iNOS2 in the WAT of FABP4^−/− mice markedly decreased postinfection. Collectively, these results suggest that C. pneumoniae lung infection in mice induces WAT pathology, characterized by robust lipolysis and the polarization of inflammatory M1-like ATMs, both of which occur in a FABP4-dependent manner (Fig. 7).

To gain insight of WAT pathology induced by C. pneumoniae and systemic metabolic disturbances, we examined the effect of acute C. pneumoniae infection on systemic glucose metabolism and insulin action in WT or FABP4^−/− mice fed a regular diet. WT mice infected with 2 × 10⁵ IFU of C. pneumoniae had a significantly impaired glucose clearance during a GTT compared with infected FABP4^−/− mice (Supplemental Fig. 4A). Comparison of the area under the curve of an oral GTT revealed that C. pneumoniae infection (2 × 10⁵ IFU) induces glucose intolerance in WT mice and this infection-induced glucose intolerance was significantly abrogated in FABP4^−/− mice, although the high-titer infection of C. pneumoniae (1 × 10⁶ IFU) led to glucose intolerance both in WT mice and FABP4^−/− mice (Supplemental Fig. 4B). Similarly, the area under the curve of insulin tolerance tests showed a significant decreased in insulin sensitivity in infected WT mice compared with FABP4^−/− mice (Supplemental Fig. 4C).

Furthermore, C. pneumoniae infection significantly increased HOMA-IR in WT mice compared with the noninfected WT mice. Importantly, azoramide treatment or 4-PBA significantly suppressed the increase in HOMA-IR compared with vehicle treatment 7 d after C. pneumoniae infection (Supplemental Fig. 4D). Collectively, these results suggest that intranasally C. pneumoniae lung infection induces WAT pathology by ER stress and impairs systemic glucose homeostasis.

We revealed in our study that C. pneumoniae lung infection induces lipolysis, promotes the accumulation of TNF-α– and IL-6–producing M1-like macrophages, and drives inflammation in WAT. Genetic ablation or chemical inhibition of FABP4 protects against C. pneumoniae–induced lipolysis and inflammation of WAT. Mechanistically, infection-induced WAT pathology is induced by ER stress/UPR, which is further amplified by FABP4 released from infected or lipolytic adipocytes and inflammatory macrophages. Treatment with azoramide, a modulator of the UPR, decreased C. pneumoniae infection–induced WAT pathology. We demonstrated in our study that acute C. pneumoniae infection in mice induces FABP4-dependent WAT pathology and provides a solid foundation for the causal relationship between C. pneumoniae infection and metabolic syndrome.

One of the notable findings of this study is that C. pneumoniae lung infection causes robust lipolysis in WAT, showing a striking dependence on FABP4. C. pneumoniae infection–induced WAT lipolysis was decreased in FABP4^−/− mice or FABP4 inhibitor–pretreated WT mice. What is the underlying mechanism of FABP4-dependent infection-induced WAT lipolysis? We previously reported that C. pneumoniae infection–induced ER stress causes lipolysis and FABP4 secretion in murine adipocytes. Exogenous FABP4 induced ER stress in HepG2 liver cells (39). Dou et al. (36) recently reported that exogenous FABP4 promotes adipocyte lipolysis and inflammation. We also observed that exogenous FABP4 treatment in differentiated 3T3L1 adipocytes induced ER stress, followed by lipolysis. Thus, we favor the hypothesis that FABP4 or fabkin secreted from C. pneumoniae–infected adipocytes may be taken up by neighboring intact adipocytes or ATMs. They may induce ER stress activation and trigger lipolysis and inflammation, followed by FABP4 secretion, leading to WAT pathology and dysfunction, ultimately resulting in the disruption of metabolic homeostasis (Fig. 7). We found that treatment with azoramide, a modulator of the ER stress/UPR, partially reversed the phenotype in mice infected with C. pneumoniae and partially canceled the WAT lipolysis (e.g., the decrease in adipocyte cell sizes, WAT weight loss), supporting the above hypothesis. It has been well appreciated that mitigation of ER stress is protective against atherosclerosis, and FABP4 is a crucial mediator of ER stress (28). Thus, our current findings of C. pneumoniae infection–induced WAT pathology augmented by FABP4 and ER stress provide a solid foundation for the causal relationship between C. pneumoniae infection and metabolic syndrome.

Another important finding in this study is the M1 macrophage polarization of ATMs postinfection with C. pneumoniae, which occurs in a FABP4-dependent manner. In the case of high-fat diet–induced adipose tissue, ER stress/UPR in macrophages, such as the activation of IRE1α and X-box binding protein 1s (XBP1s) production, has been shown to suppress M2 polarization of the ATMs while promoting proinflammatory M1 polarization (40). Yang et al. (41) reported that palmitic acid–mediated ER stress activation could induce M1-type polarization in macrophages. We have previously reported that C. pneumoniae infection in adipocytes induces ER stress and causes lipolysis and FABP4 secretion. Because our immunohistochemical investigations showed chlamydial inclusion bodies localized in ATMs and adipocytes, we speculate that C. pneumoniae infection in macrophages may induce ER stress, which could induce M1-type polarization in macrophages. Furthermore, ATMs that take up exogenous FABP4 secreted from infected adipocytes might then drive the M1 polarization of macrophages (via the induction of ER stress) (39). Therefore, it is conceivable that infection-induced ER stress and the impact of exogenous FABP4 secreted from adipocytes on ATMs and other neighboring intact adipocytes may have a fundamental role in causing WAT pathology in C. pneumoniae–infected mice (Fig. 7).

FABP4 dependency on C. pneumoniae infection–induced WAT pathology may be an essential key clue linking C. pneumoniae infection and systemic metabolic disturbances, such as hepatic steatosis and insulin resistance. The role of FABP4 in high-fat diet–induced metabolic disturbance has been well established (17). Still, the role or in vivo relevance of infectious pathogens on WAT pathology and subsequent immunometabolic diseases remains largely unknown. Further studies are warranted to reveal evidence of metabolic disturbances induced by C. pneumoniae infection–induced WAT pathology.

In summary, to our knowledge, this is the first study to describe that, other than in the context of obesity, infectious pathogens can dictate FABP4-dependent WAT pathology. A better understanding of the role of FABP4 in C. pneumoniae infection–induced WAT pathology will provide the basis for rational intervention measures directed at C. pneumoniae infection and metabolic syndrome, such as atherosclerosis, for which solid epidemiologic evidence exists.

The Hotamisligil laboratory has intellectual property related to targeting of secreted FABP4 and receives funding from Lab1636, LLC, an affiliate of Deerfield Management. G.S.H. is a scientific advisor to Crescenta BioSciences for developing small-molecule FABP inhibitors and holds equity. The other authors have no financial conflicts of interest.

We thank Bristol Myers Squibb for providing us with BMS 309403.