Maternal Adiponectin Controls Milk Composition to Prevent ...Maternal Adiponectin Controls Milk Composition to Prevent Neonatal Inflammation Zixue Jin 1, Yang Du , Adam G. Schwaid3,

Maternal Adiponectin Controls Milk Composition toPrevent Neonatal Inflammation

Zixue Jin1, Yang Du1, Adam G. Schwaid3, Ingrid W. Asterholm2,Philipp E. Scherer2, Alan Saghatelian3, and Yihong Wan1

1Department of Pharmacology, 2Touchstone Diabetes Center, Department of Internal Medicine, TheUniversity of Texas Southwestern Medical Center, Dallas, TX 75390, USA; 3Department of Chemistryand Chemical Biology, Harvard University, Cambridge, MA 02138, USA

Adiponectin is an important adipokine. Increasing evidence suggest that altered adiponectin levelsare linked with metabolic and inflammatory disorders. Here we report an important yet previouslyunrecognized function of adiponectin in lactation by which maternal adiponectin determines theinflammatory status in the nursing neonates. Surprisingly, both maternal adiponectin over-ex-pression in the transgenic mice and maternal adiponectin deletion in the knockout mice lead tosystemic inflammation in the pups, manifested as transient hair loss. However, distinct mechanismsare involved. Adiponectin deficiency triggers leukocyte infiltration and production of inflamma-tory cytokines in the lactating mammary gland. In contrast, adiponectin overabundance increaseslipid accumulation in the lactating mammary gland, resulting in excessive long chain saturatedfatty acids (LcSFA) in milk. Interestingly, in both cases, the inflammation and alopecia in the pupscan be rescued by TLR2/4 deletion because TLR2/4 double knockout pups are resistant. Mechanis-tically, LcSFA activation of inflammatory genes is TLR2/4-dependent and can be potentiated bypro-inflammatory cytokines, indicating that the inflammatory stimuli in both scenarios function-ally converge by activating the TLR2/4 signaling. Therefore, our findings reveal adiponectin as adosage-dependent regulator of lactation homeostasis and milk quality that critically control in-flammation in the nursing neonates. Furthermore, these results suggest that inflammatory infan-tile disorders may result from maternal adiponectin dysregulation that can be treated by TLR2/4inhibition.

Adiponectin is an adipocyte-derived hormone (1). It isa key modulator of lipid and glucose metabolism, and

its dysfunction is associated with many metabolic disor-ders such as type 2 diabetes, insulin resistance and ath-erosclerosis (2, 3). It is also a modulator of the innateimmunity, and its dysregulation is associated with chronicsystemic inflammation, which can further exacerbate themetabolic syndrome (4, 5). However, the role of adiponec-tin in inflammation is unclear with seemingly contradic-tory findings reported (3). Adiponectin is suggested tohave anti-inflammatory functions by stimulating a cera-midase activity because decreased adiponectin levels cor-relates with chronic inflammation associated with obesity,type 2 diabetes and cardiovascular diseases (3, 6). None-

theless, adiponectin is also proposed to be proinflamma-tory because elevated adiponectin levels correlate withclassic inflammatory and autoimmune diseases such asrheumatoid arthritis (RA) (7, 8), systemic lupus erythem-atosus (SLE) (9), inflammatory bowel disease (10) andtype 1 diabetes (11). Therefore, further investigation isrequired to elucidate the mechanisms for how adiponectincontrols inflammation (12).

Milk is the perfect food for all newborn mammals.However, our recent findings brought an unexpected twistthat maternal genetic or dietary defects, such as PPAR�/VLDLR deficiency or high-fat-diet, can lead to the secre-tion of toxic milk that causes systemic inflammation in theneonates manifested as transient alopecia (13–15). Con-

ISSN Print 0013-7227 ISSN Online 1945-7170Printed in U.S.A.Copyright © 2015 by the Endocrine SocietyReceived September 4, 2014. Accepted January 9, 2015.

Abbreviations:

R E P R O D U C T I O N - D E V E L O P M E N T

doi: 10.1210/en.2014-1738 Endocrinology endo.endojournals.org 1

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 16 January 2015. at 08:18 For personal use only. No other uses without permission. . All rights reserved.

ceptually, our findings provide new insights to the etiologyand the treatment of infantile metabolic and inflammatorydisorders. Technically, by using alopecia as a visual read-out that serves as the “eye of Drosophila”, our studieshighlight the milk-neonate axis as an innovative in vivoexperimental system to identify new regulators of metab-olism and immunity.

Here we reveal that both maternal adiponectin defi-ciency and maternal adiponectin overabundance, via dis-tinct mechanisms, cause lactation defects and milk disor-ders that lead to systemic inflammation and alopecia in thenursing neonates. These findings will enhance our under-standing of how adiponectin modulates metabolism andimmunity.

Materials and Methods

MiceAdip-KO mice and Adip-TG (ap2-�Gly-adiponectin trans-

genic) mice on a C57BL6 background have been described (16,17). TLR2/4 DKO mice on C57BL6 background have been de-scribed (18, 19). Mice were fed ad libitum with irradiated stan-dard chow (Harlan Laboratories). For mammary gland, milkand pup analyses, 8–10 week old mice were bred, and the littersizes were normalized to six pups. Milk was collected as de-scribed (13–15). For cross-fostering, pups were switched onpostnatal day 1. All protocols for mouse experiments were ap-proved by the Institutional Animal Care and Use Committee ofUniversity of Texas Southwestern Medical Center.

Histology and stainingFor immuno-fluorescence staining, skin and mammary gland

were collected and immediately embedded in OCT compound ina dry ice/ethanol bath and stored at –80°C until cryo-sectioning.Frozen OCT sections were fixed in acetone for 10 minutes andstained with FITC-conjugated antibodies specific for CD11b/Mac-1� (BD Biosciences). The sections were washed twice andmounted with medium containing DAPI (Vector Laboratories).For Oil Red O staining, frozen OCT sections were fixed in for-malin for 10 minutes, dipped in 60% isopropanol and thenstained in Oil Red O working solution (60% 3 mg/ml Oil Red Oin isopropanol) for 20 minutes. For H&E staining, tissues wereimmediately fixed in 4% paraformaldehyde (PFA), then pro-cessed, paraffin-embedded, sectioned, and H&E-stained accord-ing to standard protocols. For whole mount staining, mammaryglands were fixed in 4% PFA overnight at 4°C, rinsed in PBS,stained in carmine alum solution overnight at room temperature,and then washed in 70%, 95% and 100% ethanol each for 1hour. The tissues were cleared in xylene overnight before imageacquisition.

Gene Expression AnalysesTissue samples were snap frozen in liquid nitrogen and stored

at –80°C. Anti-TNF� antibody was from Santa Cruz Biotech-nology. Bone marrow cells were differentiated into macrophageswith 20ng/ml M-CSF for 6 days, primed with 100ng/ml TNF-�

and 100ng/ml IL-6 or PBS control for 5 hours, and then treatedwith 400 �M palmitic acid for 15 hours as described (20). RNAwas extracted using Trizol (Life Technologies), treated withRNase-free DNase I, reverse transcribed into cDNA using an ABIHigh Capacity cDNA RT Kit, and analyzed using real-timequantitative PCR (SYBR Green) in triplicate. All RNA expres-sion was normalized by two reference genes L19 and GAPDH,and similar results were observed; representative data for L19normalization are shown as relative mRNA expression. Primersequences and amplicon sizes are listed below.

Lipid Analysis by Mass SpectrometryOrganic soluble metabolites were extracted from milk with a

2:1:1 CHCl3/MeOH/H2O solution and homogenization as de-scribed (13–15). Samples were stored at –80°C and dissolved inCHCl3 prior to analysis. Ceramides were quantified by LC-MS/MS in positive ionization multiple reaction monitoring(MRM) mode as described (14). Shotgun lipidomic analyseswere performed using an Agilent Triple Quadrupole mass spec-trometer under neutral loss mode as described (14).

Statistical AnalysesAll statistical analyses were performed with Student’s t-Test

and represented as mean � standard deviation (s.d.); *, P � .05;**, P � .01; ***, P � .005; ****, P � .001; n.s. nonsignificant(P � .05).

Results and Discussion

Pups nursed by Adiponectin-KO or Adiponectin-TGdam exhibit alopecia

We found that the pups born from Adiponectin-KO(Adip-KO) or Adiponectin-TG (Adip-TG) dams started tolose dorsal hair beginning postnatal day 20–25, whichwas typically recovered one week after weaning; whereaspups born from WT littermate control mice were normal(Figure 1A-B). These pups also exhibited lower bodyweight, indicating growth retardation (Figure 1C-D). Todetermine whether the alopecia and growth retardationwas caused by maternal or offspring defect, we performedcross-fostering experiment. The results showed that thephenotype is dependent on maternal genotype because itis present in WT pups nursed by mutant dam but absent inmutant pups fostered by WT dam (Figure 1E-F). Althoughevery litter from every Adip-KO or Adip-TG dam pro-duced alopecia phenotype, on average, 78% of all pupsnursed by Adip-KO dam and 62% of all pups nursed byAdipoQ-TG dam exhibited hair loss (Figure 1E-F). Theseresults indicate that maternal adiponectin deficiency oroverabundance causes postnatal abnormality in the neo-nates manifested as hair loss and growth retardation.

Pups nursed by Adip-KO or Adip-TG dam exhibitskin inflammation

Histological analysis by H&E staining revealed that theskin of the pups with alopecia exhibited hyperplasia and

2 Dual Roles of Adiponectin in Lactation Endocrinology


follicular cysts formation (Figure 2A). Based on our pre-vious studies (13–15, 20), we examined whether the alo-pecia in the suckling pups was caused by inflammation.Immunofluorescence staining using a macrophage markerCD11b reveals that there was an increased leukocyte in-filtration in the skin of the pups nursed by either Adip-KOdam or Adip-TG dam compared to control pups nursed byWT dam (Figure 2B). Moreover, in both cases, the ex-pression of inflammatory markers was significantly ele-

vated in both skin and intestine of the defective pups (Fig-ure 2C-D). These results indicate that maternaladiponectin deficiency or overabundance can both causesystemic inflammation in the nursing neonates.

Adip-KO and Adip-TG dams exhibit distinct defectsin the lactating mammary gland

To determine the lactation defects, we performed H&Estaining, CD11b immunofluorescence staining, and Oil

Figure 1. Maternal adiponectin deletion or overabundance causes neonatal alopecia. (A-B) Mouse pups born and raised by adip-KO dam (A) oradip-Tg dam (B) developed alopecia. Representative images on postnatal day 20 (P20) are shown. C–F, Cross-fostering experiments showed thatthe alopecia in the pups is dependent on the genotype of nursing dams rather than birth dams, thus resulting from postnatal lactation defectsrather than prenatal defects. Pups were switched on postnatal day 1. C–D, Pup body weight on P20. C, Pups nursed by Adip-KO dams (n � 18)exhibited growth retardation compared to pups nursed by WT dams (n � 11). D, Pups nursed by Adip-Tg dams (n � 21) exhibited growthretardation compared to pups nursed by WT dams (n � 20). E–F, Representative images (top) and penetrance (bottom) of pup alopeciaphenotype. E, Pups nursed by Adip-KO dams exhibited alopecia but not for pups nursed by WT dams. F, Pups nursed by Adip-Tg dams exhibitedalopecia but rarely for pups nursed by WT dams.

doi: 10.1210/en.2014-1738 endo.endojournals.org 3


Red O staining of the mammary gland collected on lac-tation day 10. The results showed that there was an in-creased leukocyte infiltration in the Adip-KO gland butnot in the Adip-TG gland; in contrast, there was moreadipocytes, more and larger lipid droplets in the mammaryepithelial cells, as well as excessive lipid accumulation in

the Adip-TG gland but not in the Adip-KO gland (Figure3A). These observations suggest that the inflammation inthe pups nursed by Adip-KO and Adip-TG dams may betriggered by distinct stimuli from the milk.

We also observed different developmental defects in thelactating mammary gland from Adip-KO and Adip-TG

Figure 2. Pup hair loss is caused by follicular cysts and inflammation in the skin. (A) H&E staining showed skin hyperplasia and follicular cysts (FC,indicated by arrows) in P20 pups nursed by adip-KO or adip-Tg dams. Scale bars, 100 �m. B–E, Hair loss in pups nursed by adip-KO or adip-Tgdams was associated with a systemic inflammatory response. B, Immunofluorescence staining showed an increased infiltration of CD11b�

leukocytes in the skin. Scale bars, 25 �m. C–D, RT-QPCR analysis showed an increased expression of proinflammatory cytokines in the skin (C) andintestine (D) (n � 6). Pups of the same genotype were compared.



Figure 3. Adiponectin deletion or overabundance causes distinct changes in the lactating mammary gland. (A) Lactating mammary glands fromAdip-KO dams exhibited increased leukocyte infiltration, whereas lactating mammary glands from Adip-Tg dams exhibited increased lipidaccumulation. Mammary glands were collected on lactation day 10 and subjected to H&E, CD11b immunofluorescence, or Oil Red O staining.Green arrows indicate leukocytes. Red arrows indicate adipocytes. Blue arrows indicate larger lipid droplets in the lactating mammary epitheliumof Adip-Tg dams. Scale bars, 25 �m. B, Whole mount staining showed a reduced alveolar expansion in P15 (pregnancy day 15) mammary glandfrom Adip-Tg dams but not Adip-KO dams. Scale bars, 60 �m. Histomorphometric quantification of lobuloalveolar area (%) and alveoli counts pergland is shown (n � 6). C, RT-QPCR analysis showed that involution-associated proapoptotic genes were increased whereas antiapoptotic geneswere decreased in L10 Adip-KO glands but not in L10 Adip-Tg glands (n � 6). Gene expression relative to L19 (top) or GAPDH (bottom) showed



dams. Adip-TG glands exhibited a reduced alveolar ex-pansion during pregnancy and lactation, resulting in moreadipocytes (Figure 3A-B). Adip-KO glands exhibited anearly involution, leading to more apoptosis and less duc-tal/alveolar structures in the L10 mammary glands (Figure3A,C). These developmental defects can compromise ef-ficient lactation and reduce milk production, which is con-sistent with the lower body weight in the pups nursed byAdip-KO and Adip-TG dams (Figure 1C-D).

Adip-KO dams produce more proinflammatorycytokines in the lactating mammary gland

Consistent with the increased leukocyte infiltration inthe Adip-KO lactating mammary gland (Figure 3A), theexpression of many proinflammatory cytokines was sig-nificantly elevated, including IL-6, TNF-�, IL-1�, MMP9,IL-12P35, IL-12P40 (Figure 4A). Moreover, the expres-sion of several enzymes that produce proinflammatorylipids such as prostaglandin and leukotriene was alsohigher, including COX-2, Alox-5 and Ephx1 (Figure 4B).Once again, these proinflammatory changes were eitherabsent or less severe in the Adip-TG glands (Figure 4A-B).The expression of adiponectin in the lactating mammarygland was ablated in Adip-KO mice but elevated by 3.4fold in Adip-TG mice, whereas the expression of adi-ponectin receptors was unaltered (Figure 4C). In accor-dance with the elevated mRNA expression of proinflam-matory cytokines (Figure 4A) and lipid oxidation enzymesin the mammary gland of Adip-KO mice (Figure 4B), theprotein levels of TNF� was higher in the milk fromAdip-KO dams (Figure 4D) and the levels of oxidized fattyacids were increased in the skin of the pups nursed byAdip-KO dams (Figure 4E). These data suggest that thealopecia and inflammation in the pups nursed by Adip-KOdams may be triggered by the excessive inflammatory cy-tokines in the milk.

Adip-TG dams produce more long chain saturatedfatty acids in the milk

Previous studies report that adiponectin overabun-dance in the Adip-TG mice promotes the clearance of trig-lycerides and fatty acids (FAs) from circulation (16, 21).Thus, we hypothesize that more fat may be transferred tothe lactating mammary gland and secreted into milk in theAdip-TG dams. Consistent with the increased lipid accu-mulation and larger milk lipid droplets in the Adip-TGlactating mammary gland (Figure 3A), the lipid content inthe milk was also significantly higher in Adip-TG micecompared to WT controls or Adip-KO mice (Figure 4F).

Triglycerides make up 98% of milk lipids, and their FAsderive from both de novo synthesis in the lactating mam-mary gland which generates predominantly short and me-dium chain (C8-C12) FAs, as well as adipose tissue andcirculation that contains mainly long chain (C16-C20)FAs (22). Previous studies show that inflammation is spe-cifically activated by saturated FAs but not unsaturatedFAs (23–27); and by long chain FAs (C16:0, C18:0) butnot medium/short chain (C8:0-C12:0) FAs (28). There-fore, we investigated the FA composition of milk triglyc-erides by shotgun lipidomic analysis using neutral lossmass spectrometry (29). The results showed that the FAelongation ratio (C18:0/C10:0) (Figure 4G) was 3.6-foldhigher in the Adip-TG milk, indicating a higher percentageof long chain FAs. Moreover, the FA saturation ratio(C18:0/C18:2) was also 1.9-fold higher in the Adip-TGmilk (Figure 4H), indicating a higher percentage of satu-rated FAs. Particularly, the level of palmitic acid (C16:0),a representative long-chain saturated fatty acid (LcSFA)that induces inflammation via TLR4 activation (27, 30–34), was 2.3-fold higher in the Adip-TG milk compared toWT milk (Figure 4I). In contrast, these changes were notobserved for Adip-KO milks (Figure 4G-I).

Saturated FAs induce the biosynthesis of ceramides(Cer) and glucosylceramides (GlcCer) to promote inflam-mation and metabolic diseases (35, 36). We found thatthere were also higher levels of Cer and GlcCer in theAdip-TG milk (Figure 4J). This is a local effect in the lac-tating mammary gland as systemic ceramide levels werelower in the Adip-TG mice (6). Together, these resultssuggest that the alopecia and inflammation in the pupsnursed by Adip-TG dams may be triggered by the excessiveLcSFAs and ceramides in the milk.

TLR2/4 DKO pups are resistant to alopecia whennursed by Adip-KO or Adip-TG dams

TLR2/4 signaling is an important nexus of inflamma-tory response pathways. TLR2/4 can be activated byLcSFAs such as palmitic acid (PA) (27, 30–34), which areexcessive in Adip-TG milk (Figure 4G-I). TLR2/4 functioncan also be primed and synergized by proinflammatorycytokines (37–39), which are excessive in Adip-KO milk.We found that macrophage activation by palmitic acid, ahighly abundant FA in milk, was significantly augmentedby TNF� and IL-6, two cytokines that were elevated in theAdip-KO lactating mammary gland (Figure 4K).

We next investigated whether the neonatal inflamma-tion triggered by the toxic milk produced by Adip-KO and

Legend to Figure 3 Continued. . .similar results.



Figure 4. Increased inflammatory cytokines in Adip-KO mammary glands and excessive long-chain saturated fatty acids in Adip-Tg milk. (A)Expression of proinflammatory cytokines was increased in the lactating mammary glands from Adip-KO dam but not from Adip-Tg dams (n � 6).B, Expression of enzymes that synthesize proinflammatory lipids was increased in the lactating mammary glands from Adip-KO dam but not fromAdip-Tg dams (n � 6). C, Expression of adiponectin, adiponectin receptor 1 (Adip-R1) and adiponectin receptor 2 (Adip-R2) in the lactatingmammary glands (n � 6). D, Western blot showed that milk from Adip-KO dams, but not from Adip-Tg dams, had increased TNF� protein levels.Equal volume (15�l) of milk samples were loaded. E, LC-MS analysis showed that levels of hydroxyl fatty acids were increased in the skin of thepups nursed by Adip-KO dams but not Adip-Tg dams (n � 6). F, Milk from Adip-Tg dams, but not from Adip-KO dams, had increased lipid content(n � 3). G–J, LC-MS analyses of milk lipids. G, Milk from Adip-Tg dams, but not Adip-KO dams, showed increased fatty acid elongation ratio,indicating a higher percentage of long chain fatty acids (n � 3). H, Milk from Adip-Tg dams, but not Adip-KO dams, showed increased fatty acidsaturation ratio, indicating a higher percentage of saturated fatty acids (n � 3). I, Milk from Adip-Tg dams, but not Adip-KO dams, contained



Adip-TG dams can be rescued by TLR2/4 deletion. Whenfostered by Adip-KO or Adip-TG dams, while WT controlpups still developed hair loss, TLR2/4 double knockout(DKO) pups were completely resistant (Figure 5A-B);while the expression of inflammatory genes was still ele-vated in WT control pups, this increase was largely abol-ished in TLR2/4 DKO pups (Figure 5C-D). Interestingly,the lower body weight in the pups nursed by Adip-KO orAdip-Tg dams was not significantly rescued by TLR2/4deletion (Figure 5A-B), suggesting that TLR2/4 deletion

specifically abolished the systemic inflammation causedby altered milk composition but not the growth retarda-tion caused by mammary gland developmental defects.Consistent with these observations, palmitic acid induc-tion and cytokine potentiation of inflammatory geneswere attenuated in TLR2/4DKO macrophages (Figure5E). These results indicate that, despite the distinct in-flammatory stimuli from Adip-KO and Adip-TG milk, theinflammatory signaling converges at TLR2/4 activation in

Legend to Figure 4 Continued. . .higher levels of palmitic acid (PA) (n � 3). J, Milk from Adip-Tg dams, but not Adip-KO dams, contained higher levels of ceramides (Cer) andglucosylceramide (GlcCer) (n � 3). K, Palmitic acid (PA) induction of inflammatory genes was potentiated by cytokines. Bone marrow cells weredifferentiated into macrophages with 20ng/ml M-CSF for 6 days, primed with 100ng/ml TNF-� and 100ng/ml IL-6 or PBS control for 5 hours, andthen treated with 400 �M palmitic acid (PA) for 15 hours before gene expression analysis (n � 3).

Figure 5. TLR2/4 deletion in the pups rescues the inflammation and hair loss. (A) TLR2/4 DKO pups, when fostered by Adip-KO dams, wereresistant to alopecia. B, TLR2/4 DKO pups, when fostered by Adip-Tg dams, were resistant to alopecia. C, TLR2/4 DKO pups, when fostered byAdip-KO dams, were resistant to the increased inflammatory gene expression in the skin. D, TLR2/4 DKO pups, when fostered by Adip-Tg dams,were resistant to the increased inflammatory gene expression in the skin. E, Palmitic acid (PA) induction and cytokine potentiation of inflammatorygenes were attenuated in TLR2/4 DKO macrophages. Bone marrow cells were differentiated into macrophages with 20ng/ml M-CSF for 6 days,primed with 100ng/ml TNF-� and 100ng/ml IL-6 or PBS control for 5 hours, and then treated with 400 �M palmitic acid (PA) for 15 hours beforegene expression analysis (n � 3).



the nursing neonates and can be prevented by TLR2/4inhibition.

Our study reveals maternal adiponectin as a key geneticprogram that controls metabolic and immune homeosta-sis during lactation to ensure normal milk compositionand prevent neonatal inflammation. Interestingly, thisregulation is dosage-dependent because adverse effectscan result from either adiponectin deficiency or adiponec-tin overabundance. On one hand, lack of maternal adi-ponectin causes increased production of inflammatory cy-tokines in the lactating mammary gland, which in turnprimes and synergizes fatty acid activation of TLR2/4 sig-naling pathway in the neonates, leading to an overt in-flammatory response. On the other hand, excessive ma-ternal adiponectin promotes triglycerides accumulation inthe lactating mammary gland, causing higher levels oflong-chain saturated fatty acids in milk, which in turnactivate TLR2/4 to mount an inflammatory response inthe newborns (Figure 6). Thus, adiponectin control of in-flammation involves the regulation of both metabolismand immunity. These findings demonstrate that physio-logical level of adiponectin must be within an optimalwindow to prevent inflammation, providing insights tothe perplexing and seemingly contradictory observationsof its both antiand proinflammatory effects in experimen-tal and clinical investigations.

Changes of the lipid or cytokine content in the milk mayalso alter the microbiome of the neonatal gut, which inturn may modulate local immune cell function and/or nu-trient uptake, and consequently contribute to the systemicinflammation phenotype in the offspring. In future stud-ies, it would be interesting to compare the major gut phylain the pups nursed by Adip-KO or Adip-Tg dams vs. pupsnursed by WT dams.

The etiology of several inflammatory infantile disor-

ders is still unclear, such as asthma, eczema, inflammatorybowel disease, necrotizing enterocolitis (NEC) and neo-natal onset multisystem inflammatory disease. Our find-ings suggest that these diseases may originate from milkdisorders caused by maternal dysfunction of adiponectin,in addition to our previously identified genetic or dietaryfactors (13–15). Importantly, we found that the inflam-matory insults in both maternal adiponectin deficiencyand maternal adiponectin overabundance converge toTLR2/4 activation, and TLR2/4 deletion in the offspringconfers resistance. Therefore, our findings reveal thatTLR2/4 inhibition may represent a novel strategy for theprevention and treatment of infantile inflammatorydisorders.

Acknowledgments

We thank UTSW Molecular Pathology Core for histology. Y.W.is a Virginia Murchison Linthicum Scholar in Medical Research.This work was in part supported by March of Dimes (#6-FY13–137, YW), The Welch Foundation (I-1751, YW), NIH (R01DK089113, YW), CPRIT (RP130145, YW), DOD BCRP IdeaAward (W81XWH-13–1-0318, YW) and UTSW EndowedScholar Startup Fund (YW). The authors declare that they haveno financial conflict of interest.

Address all correspondence and requests for reprints to: As-sistant Professor Yihong Wan, Ph.D., Email:[email protected], Univ of Texas SouthwesternMedical Center - Department of Pharmacology, 6001 ForestPark Road, Rm ND8.502B, Dallas, TX, UNITED STATES,75390–9041, Phone: 214–645-6062

This work was supported by .

Table 1.

MouseGene Primer-F Primer-R

Amplicon(bp)

IL-1b TGGGCCTCAAAGGAAAGAATC GGTATTGCTTGGGATCCACACT 93MMP9 CCAAGGGTACAGCCTGTTCCT GCACGCTGGAATGATCTAAGC 74IL-6 TCCCCATCTCTCATGCAGTGT CTCTCTCCCTTCTGAGCAGCTG 90TNF-� CAGCCGATGGGTTGTACCTT GTGTGGGTGAGGAGCACGTA 82IL-12P35 AAATGAAGCTCTGCATCCTGCT AGATAGCCCATCACCCTGTTGA 73IL-12P40 CATCTACCGAAGTCCAATGCAA AGGGAACACATGCCCACTTG 96COX-2 ATCAGAACCGCATTGCCTCT CCAGGAGGATGGAGTTGTTGTAG 129Ephx1 TGTGCCCACTGGCTATTCAG GGGTACTTGACCTTCACCCACTT 75Alox5 GTGCCTCTTCCAAGCTCGAA CCCATCCTACAACAGCCTTCA 76Adiponectin CAGTGGATCTGACGACACCAA GAACAGGAGAGCTTGCAACAGT 59Adip-R1 TCTCCATCGTCTGTGTCCTG AGTGCATGGTGGGTACAACA 138Adip-R2 GTGGAGCTCAGCAGTACCCT TGTCCCTTCTTCTTGGGAGA 102L19 GGTCTGGTTGGATCCCAATG CCCATCCTTGATCAGCTTCCT 82GAPDH ACCTGCCAAGTATGATGACATCA CCCTCAGATGCCTGCTTCAC 51



mailto:[email protected]

References

1. Scherer PE, Williams S, Fogliano M, Baldini G, Lodish HF. A novelserum protein similar to C1q, produced exclusively in adipocytes.The Journal of biological chemistry. 1995;270(45):26746–26749.

2. Berg AH, Combs TP, Scherer PE. ACRP30/adiponectin: an adipo-kine regulating glucose and lipid metabolism. Trends in endocrinol-ogy and metabolism: TEM. 2002;13(2):84–89.

3. Ouchi N, Parker JL, Lugus JJ, Walsh K. Adipokines in inflammationand metabolic disease. Nature reviews Immunology. 2011;11(2):85–97.

4. Hotamisligil GS, Erbay E. Nutrient sensing and inflammation inmetabolic diseases. Nature reviews Immunology. 2008;8(12):923–934.

5. Hotamisligil GS. Inflammation and metabolic disorders. Nature.2006;444(7121):860–867.

6. Holland WL, Miller RA, Wang ZV, Sun K, Barth BM, Bui HH,Davis KE, Bikman BT, Halberg N, Rutkowski JM, Wade MR,Tenorio VM, Kuo MS, Brozinick JT, Zhang BB, Birnbaum MJ,Summers SA, Scherer PE. Receptor-mediated activation of cerami-dase activity initiates the pleiotropic actions of adiponectin. NatMed. 2011;17(1):55–63.

7. Schaffler A, Ehling A, Neumann E, Herfarth H, Tarner I, Schol-merich J, Muller-Ladner U, Gay S. Adipocytokines in synovial fluid.JAMA : the journal of the American Medical Association. 2003;290(13):1709–1710.

8. Otero M, Lago R, Gomez R, Lago F, Dieguez C, Gomez-Reino JJ,Gualillo O. Changes in plasma levels of fat-derived hormones adi-ponectin, leptin, resistin and visfatin in patients with rheumatoidarthritis. Annals of the rheumatic diseases. 2006;65(9):1198–1201.

9. Sada KE, Yamasaki Y, Maruyama M, Sugiyama H, Yamamura M,Maeshima Y, Makino H. Altered levels of adipocytokines in asso-

ciation with insulin resistance in patients with systemic lupus ery-thematosus. The Journal of rheumatology. 2006;33(8):1545–1552.

10. Fayad R, Pini M, Sennello JA, Cabay RJ, Chan L, Xu A, Fantuzzi G.Adiponectin deficiency protects mice from chemically induced co-lonic inflammation. Gastroenterology. 2007;132(2):601–614.

11. Heliovaara MK, Teppo AM, Karonen SL, Tuominen JA, Ebeling P.Improved glycaemia in type 1 diabetes results in decreased levels ofsoluble adhesion molecules with no change in serum adiponectin ormost acute phase proteins. Experimental and clinical endocrinol-ogy, diabetes : official journal, German Society of Endocrinology[and] German Diabetes Association. 2006;114(6):295–300.

12. Fantuzzi G. Adiponectin and inflammation: consensus and contro-versy. The Journal of allergy and clinical immunology. 2008;121(2):326–330.

13. Wan Y, Saghatelian A, Chong LW, Zhang CL, Cravatt BF, EvansRM. Maternal PPAR gamma protects nursing neonates by suppress-ing the production of inflammatory milk. Genes Dev. 2007;21(15):1895–1908.

14. Du Y, Yang M, Lee S, Behrendt CL, Hooper LV, Saghatelian A, WanY. Maternal western diet causes inflammatory milk and TLR2/4-dependent neonatal toxicity. Genes Dev. 2012;26(12):1306–1311.

15. Du Y, Yang M, Wei W, Huynh HD, Herz J, Saghatelian A, Wan Y.Macrophage VLDL receptor promotes PAFAH secretion in moth-er’s milk and suppresses systemic inflammation in nursing neonates.Nat Commun. 2012;3:1008.

16. Combs TP, Pajvani UB, Berg AH, Lin Y, Jelicks LA, Laplante M,Nawrocki AR, Rajala MW, Parlow AF, Cheeseboro L, Ding YY,Russell RG, Lindemann D, Hartley A, Baker GR, Obici S, DeshaiesY, Ludgate M, Rossetti L, Scherer PE. A transgenic mouse with adeletion in the collagenous domain of adiponectin displays elevatedcirculating adiponectin and improved insulin sensitivity. Endocri-nology. 2004;145(1):367–383.

17. Nawrocki AR, Rajala MW, Tomas E, Pa-jvani UB, Saha AK, Trumbauer ME,Pang Z, Chen AS, Ruderman NB, ChenH, Rossetti L, Scherer PE. Mice lackingadiponectin show decreased hepatic in-sulin sensitivity and reduced responsive-ness to peroxisome proliferator-activatedreceptor gamma agonists. The Journal ofbiological chemistry. 2006;281(5):2654–2660.

18. Hoshino K, Takeuchi O, Kawai T, SanjoH, Ogawa T, Takeda Y, Takeda K, AkiraS. Cutting edge: Toll-like receptor 4(TLR4)-deficient mice are hyporespon-sive to lipopolysaccharide: evidence forTLR4 as the Lps gene product. J Immu-nol. 1999;162(7):3749–3752.

19. Takeuchi O, Hoshino K, Kawai T, SanjoH,TakadaH,OgawaT,TakedaK,AkiraS. Differential roles of TLR2 and TLR4 inrecognition of gram-negative and gram-positive bacterial cell wall components.Immunity. 1999;11(4):443–451.

20. Jin Z, Wei W, Yang M, Du Y, Wan Y.Mitochondrial Complex I Activity Sup-presses Inflammation and Enhances BoneResorption by Shifting Macrophage-Os-teoclast Polarization. Cell Metab. 2014:E-pub.

21. Shetty S, Ramos-Roman MA, Cho YR,Brown J, Plutzky J, Muise ES, Horton JD,Scherer PE, Parks EJ. Enhanced fatty acidflux triggered by adiponectin overexpres-sion. Endocrinology. 2012;153(1):113–122.

Figure 6. A schematic diagram for the dosage-dependent regulation of neonatal inflammationby maternal adiponectin via distinct mechanisms. Maternal adiponectin deficiency increasesinflammatory cytokines in the mammary gland and milk, which potentiate the ability of milk fattyacids to activate inflammatory genes in a TLR2/4-dependent manner. Maternal adiponectinoverabundance increases the levels of long-chain saturated fatty acids in the milk, which activateinflammatory genes in a TLR2/4-dependent manner. Both cause a systemic inflammatoryresponse in the nursing neonates manifested as alopecia. Our findings indicate that fatty acidsactivation of TLR2/4 is an important proinflammatory component in both Adip-Tg milk (whichhas increased abundance of saturated fatty acids) and Adip-KO milk (which has increasedproinflammatory activity of saturated fatty acids due to the excessive cytokines).



22. Neville MC, Picciano MF. Regulation of milk lipid secretion andcomposition. Annu Rev Nutr. 1997;17:159–183.

23. Lee JY, Plakidas A, Lee WH, Heikkinen A, Chanmugam P, Bray G,Hwang DH. Differential modulation of Toll-like receptors by fattyacids: preferential inhibition by n-3 polyunsaturated fatty acids. JLipid Res. 2003;44(3):479–486.

24. Lee JY, Sohn KH, Rhee SH, Hwang D. Saturated fatty acids, but notunsaturated fatty acids, induce the expression of cyclooxygenase-2mediated through Toll-like receptor 4. J Biol Chem. 2001;276(20):16683–16689.

25. Lee JY, Ye J, Gao Z, Youn HS, Lee WH, Zhao L, Sizemore N,Hwang DH. Reciprocal modulation of Toll-like receptor-4 signalingpathways involving MyD88 and phosphatidylinositol 3-kinase/AKT by saturated and polyunsaturated fatty acids. J Biol Chem.2003;278(39):37041–37051.

26. Saberi M, Woods NB, de Luca C, Schenk S, Lu JC, BandyopadhyayG, Verma IM, Olefsky JM. Hematopoietic cell-specific deletion oftoll-like receptor 4 ameliorates hepatic and adipose tissue insulinresistance in high-fat-fed mice. Cell Metab. 2009;10(5):419–429.

27. Shi H, Kokoeva MV, Inouye K, Tzameli I, Yin H, Flier JS. TLR4links innate immunity and fatty acid-induced insulin resistance.J Clin Invest. 2006;116(11):3015–3025.

28. Hommelberg PP, Plat J, Langen RC, Schols AM, Mensink RP. Fattyacid-induced NF-kappaB activation and insulin resistance in skeletalmuscle are chain length dependent. Am J Physiol Endocrinol Metab.2009;296(1):E114–120.

29. Han X, Gross RW. Shotgun lipidomics: electrospray ionizationmass spectrometric analysis and quantitation of cellular lipidomesdirectly from crude extracts of biological samples. Mass SpectromRev. 2005;24(3):367–412.

30. Holland WL, Bikman BT, Wang LP, Yuguang G, Sargent KM, Bul-chand S, Knotts TA, Shui G, Clegg DJ, Wenk MR, Pagliassotti MJ,Scherer PE, Summers SA. Lipid-induced insulin resistance mediatedby the proinflammatory receptor TLR4 requires saturated fatty ac-id-induced ceramide biosynthesis in mice. The Journal of clinicalinvestigation. 2011;121(5):1858–1870.

31. Caricilli AM, Nascimento PH, Pauli JR, Tsukumo DM, Velloso LA,

Carvalheira JB, Saad MJ. Inhibition of toll-like receptor 2 expres-sion improves insulin sensitivity and signaling in muscle and whiteadipose tissue of mice fed a high-fat diet. J Endocrinol. 2008;199(3):399–406.

32. Poggi M, Bastelica D, Gual P, Iglesias MA, Gremeaux T, Knauf C,Peiretti F, Verdier M, Juhan-Vague I, Tanti JF, Burcelin R, AlessiMC. C3H/HeJ mice carrying a toll-like receptor 4 mutation areprotected against the development of insulin resistance in white ad-ipose tissue in response to a high-fat diet. Diabetologia. 2007;50(6):1267–1276.

33. Senn JJ. Toll-like receptor-2 is essential for the development ofpalmitate-induced insulin resistance in myotubes. J Biol Chem.2006;281(37):26865–26875.

34. Tsukumo DM, Carvalho-Filho MA, Carvalheira JB, Prada PO, Hi-rabara SM, Schenka AA, Araujo EP, Vassallo J, Curi R, Velloso LA,Saad MJ. Loss-of-function mutation in Toll-like receptor 4 preventsdiet-induced obesity and insulin resistance. Diabetes. 2007;56(8):1986–1998.

35. Holland WL, Brozinick JT, Wang LP, Hawkins ED, Sargent KM,Liu Y, Narra K, Hoehn KL, Knotts TA, Siesky A, Nelson DH,Karathanasis SK, Fontenot GK, Birnbaum MJ, Summers SA. Inhi-bition of ceramide synthesis ameliorates glucocorticoid-, saturated-fat-, and obesity-induced insulin resistance. Cell Metab. 2007;5(3):167–179.

36. Holland WL, Summers SA. Sphingolipids, insulin resistance, andmetabolic disease: new insights from in vivo manipulation of sph-ingolipid metabolism. Endocr Rev. 2008;29(4):381–402.

37. Mosser DM, Edwards JP. Exploring the full spectrum of macro-phage activation. Nature reviews Immunology. 2008;8(12):958–969.

38. Baker RG, Hayden MS, Ghosh S. NF-kappaB, inflammation, andmetabolic disease. Cell metabolism. 2011;13(1):11–22.

39. Bosisio D, Polentarutti N, Sironi M, Bernasconi S, Miyake K, WebbGR, Martin MU, Mantovani A, Muzio M. Stimulation of toll-likereceptor 4 expression in human mononuclear phagocytes by inter-feron-gamma: a molecular basis for priming and synergism withbacterial lipopolysaccharide. Blood. 2002;99(9):3427–3431.



Documents

Maternal Adiponectin Controls Milk Composition to Prevent ...Maternal Adiponectin Controls Milk Composition to Prevent Neonatal Inflammation Zixue Jin 1, Yang Du , Adam G. Schwaid3,