Isolation of the cDNA encoding the acid labile subunit

Isolation of the cDNA encoding the acid labile subunit (ALS) of the150 kDa IGF-binding protein complex in cattle and ALS regulationduring the transition from pregnancy to lactation

Jin W Kim, Robert P Rhoads, Nthabisheng Segoale,Niels B Kristensen1, Dale E Bauman and Yves R BoisclairDepartment of Animal Science, 259 Morrison Hall, Cornell University, Ithaca, New York 14853-4801, USA1Department of Animal Health, Welfare and Nutrition, Danish Institute of Agricultural Sciences, PO Box 50, 8830 Tjele, Denmark

(Requests for offprints should be addressed to Y R Boisclair; Email: [email protected])

Abstract

During the transition from pregnancy to lactation, dairycows experience a 70% reduction in plasma IGF-I. Thisreduction has been attributed to decreased hepatic IGF-Iproduction. IGF-I circulates predominantly in multi-protein complexes consisting of one molecule each ofIGF-I, IGF binding protein-3 and the acid labile subunit(ALS). Recent studies in the mouse have shown thatabsence of ALS results in accelerated turnover and severelydepressed concentration of plasma IGF-I. These observa-tions suggest that reduced plasma ALS could be a secondfactor contributing to the fall of plasma IGF-I in peri-parturient cows. This possibility has not been studied dueto the lack of bovine ALS reagents. To address this, weisolated the bovine ALS cDNA and used its sequence todevelop a ribonuclease protection assay (RPA) and abovine ALS antiserum. Using the RPA, ALS mRNAabundance was approximately fivefold higher in liver thanin lung, small intestine, adipose tissue, kidney and heart,but was absent in muscle and brain. The antiserumdetected the highest ALS levels in plasma followed byovarian follicular fluid, lymph and colostrum. A portionof colostrum and follicular fluid ALS appears to be

synthesized locally as ALS mRNA was found in mammaryepithelial cells and ovarian follicular cells. Finally, wemeasured plasma ALS in dairy cows during the peri-parturient period (days �35 and +56 relative to partu-rition on day 0). Plasma ALS dropped by 50% betweenlate pregnancy and the first day of lactation and returnedto prepartum levels by day +56. To determine whetherthis reflected a change in hepatic expression, ALS mRNAwas measured in liver biopsies collected on days �35, +3and +56. ALS mRNA expression was significantly loweron day +3 than on day �35, but recovered completelyby day +56. Finally, we examined the ability of GH toincrease plasma ALS abundance at selected times beforeand after parturition (weeks �5, �2, +1 and +5). GHincreased plasma ALS at weeks �5, �2 and +5, but notat week +1. Identical effects of GH were seen when theresponse considered was plasma IGF-I. We conclude thatthe decline in plasma ALS after parturition is a conse-quence of hepatic GH resistance and contributes to theassociated reduction of plasma IGF-I.Journal of Endocrinology (2006) 189, 583–593

Introduction

During the transition from late pregnancy to early lacta-tion, dairy cows enter a period of severe negative energybalance (Bell 1995, Boisclair et al. 2006). Under theseconditions, changes in the plasma concentration of keyhormones coordinate adaptations required for the preser-vation of metabolic homeostasis (Bell 1995, Boisclair et al.2006). For example, plasma insulin-like growth factor-I(IGF-I) is decreased by 70% after parturition, facilitatingthe mobilization of amino acids from skeletal muscle insupport of hepatic gluconeogenesis (Bell 1995, Boisclairet al. 2006). Based on positive correlations between

plasma IGF-I concentrations and hepatic IGF-I mRNAabundance, the periparturient drop in plasma IGF-Ihas been attributed to decreased hepatic production(Kobayashi et al. 1999, Radcliff et al. 2003, Kim et al. 2004).

Decreased hepatic production, however, is unlikely tobe the only factor involved. This is because plasma IGF-Ialmost always circulates in multi-proteins, high molecularweight complexes (Baxter 1994, Boisclair et al. 2001). Inadult animals, the predominant complex consists of onemolecule each of IGF-I, IGF binding protein-3 or -5(IGFBP-3 or -5), and the acid labile subunit (ALS) (Baxter1994, Twigg & Baxter 1998, Boisclair et al. 2001).Incorporation of free IGF-I in these ternary complexes

583

Journal of Endocrinology (2006) 189, 583–5930022–0795/06/0189–583 � 2006 Society for Endocrinology Printed in Great Britain

DOI: 10.1677/joe.1.06824Online version via http://www.endocrinology-journals.org

Downloaded from Bioscientifica.com at 11/22/2021 05:30:08PMvia free access

http://www.endocrinology-journals.org

extends its half-life from minutes to over 12 h (Guler et al.1989, Baxter 1994, Zapf et al. 1995). Adult animals devoidof ALS have accelerated turnover and severely depressedconcentration of plasma IGF-I (Ueki et al. 2000, Yakar etal. 2002, Domene et al. 2004). ALS deficiency also resultsin the near disappearance of plasma IGFBP-3 whereasabsence of IGFBP-3 has no effect on plasma ALS (Uekiet al. 2000, Yakar et al. 2005). These observations suggestthat variation in plasma ALS could be a second factorcontributing to the fall of plasma IGF-I in periparturientdairy cows.

The ALS cDNA and gene were isolated first in human,mouse and rat (Boisclair et al. 2001), and consequentlymost of the existing information on ALS was derived fromthese species. These studies have shown that ALS issynthesized primarily in liver in a growth hormone(GH)-dependent manner (Baxter 1990, Baxter & Dai1994). The effects of GH occur via a stimulation of genetranscription, resulting in increased abundance of ALSmRNA in liver and ALS in plasma (Dai & Baxter 1994,Ooi et al. 1997, Olivecrona et al. 1999). In contrast, ALShas not been studied in dairy cattle, including during thetransition from pregnancy to lactation when plasma IGF-Ivaries dynamically. The major factor contributing to thislack of information is the poor cross-reactivity of anti-bodies raised against rodent and human ALS with bovineALS (Baxter 1990, Baxter & Dai 1994).

To address the lack of information on ALS in dairycattle, we pursued two specific objectives. First, we clonedthe bovine ALS cDNA and used this sequence to developbovine-specific reagents. Secondly, using these reagents,we asked whether ALS is regulated during the transitionfrom pregnancy to lactation in dairy cattle. Our studiesshow that hepatic ALS expression and production declinein early lactating dairy cows as a consequence of hepaticGH resistance.

Materials and Methods

All experimental procedures involving animals wereapproved by the Animal Care and Use Committees ofCornell University and the Danish Animal ExperimentsInspectorate.

Cloning of the bovine ALS cDNA

The full-length bovine ALS cDNA was amplified using aforward primer corresponding to nt �57 to �38 of theovine ALS cDNA (relative to A+1TG; 5�-ATCCAGAGGGCAGGAGCAGC-3�) and a poly A tail-specific reverseprimer (5�-T30NN-3�; Clontech laboratories, Palo Alto,CA, USA). PCR was performed using a high fidelitypolymerase as recommended by the manufacturer (ExTaq polymerase; Takara Bio Inc, Japan). Products were

subcloned into the vector pCR 2·1-Topo (Invitrogen),sequenced and submitted to Genebank (accession numberDQ444712).

Spatial ALS expression

Tissues were obtained at slaughter from two adult dairycows that had been lactating for �300 days (see Fig. 2 forcomplete listing of tissues). Ovarian follicular cells wereisolated from freshly excised dominant and subdominantfollicles (�5 mm) and pooled. Mammary epithelial cellswere isolated by digesting the parenchyma of a lactatingdairy cow with a mixture of a collagenase, hyaluronidaseand elastase (Matitashvili & Bauman 1999). Cells weresubjected to gradient centrifugation to yield primarilymammary epithelial cells (>95% of isolated cells). Alltissues and cells were frozen in liquid N2 immediatelyupon isolation.

Total RNA was extracted by the guanidiumthiocyanate-phenol-chloroform method. The concen-tration of total RNA was determined by absorbance at260 nm. RNA quality was verified by electrophoresis onformaldehyde agarose gel and ribosomal bands stainingwith Sybr Green II (Molecular Probes, Eugene, OR,USA). ALS mRNA abundance was measured by ribo-nuclease protection assay (RPA) with an RPA III kit(Ambion, Austin, TX, USA). The ALS probe corre-sponded to nt +182 to nt +599 (A+1TG) of the bovineALS cDNA. The RPA was performed in the presenceof a tenfold molar excess of a low-specific-activity18S riboprobe generated from an 18S DNA template(Ambion). Signals were quantified by phosphorimagingand normalized to the 18S signal.

Production and validation of bovine ALS antisera

A peptide corresponding to the 28 carboxyl-terminalresidues of bovine ALS (ALSc28) was conjugated tokeyhole limpet hemocyanin (KLH) using sulfhydrylchemistry (Bio-Synthesis Inc, Lewisville, TX, USA).Conjugated ALSC28 (100 µg) was emulsified with an equalvolume of complete Freund’s adjuvant and injected sub-cutaneously into three New Zealand White rabbits. Forsecondary immunizations, emulsions were prepared withincomplete Freund’s adjuvant and injected on days 14, 42,63, 84 and 105 (relative to primary immunization).Rabbits were bled on day 115 and serum was storedat �20 �C.

Immunoreactivity of antisera was evaluated with plasmaobtained from adult animals of various species (listed inFig. 3B). In some studies, bovine plasma was deglyco-sylated. Briefly, plasma (2 µl) was boiled in 0·5% SDS, 1%�-mercaptoethanol for 10 min. After adding NaHPO4and 10% NP-40 to final concentrations of 0·05 M and 1%respectively, the samples were incubated at 37 �C in theabsence or presence of 500 units of peptide-N-glycosidase F

J W KIM and others · Acid labile subunit in the dairy cow584

www.endocrinology-journals.orgJournal of Endocrinology (2006) 189, 583–593



(PNGase F; New England Biolabs, Beverly, MA, USA).After 1 h, reactions were terminated by boiling in 1�Laemmli loading buffer. Native and deglycosylated plasmawere electrophoresed under reducing conditions on 10%SDS–polyacrylamide gels, and electroblotted onto nitro-cellulose membranes (Protran; Schleicher & Schuell Bio-science, Keene, NH, USA). The membranes wereblocked in Tris-buffered saline with Tween-20 (TBST;0·05 M Tris, pH 7·4, 0·2 M NaCl, 0·1% Tween-20)containing 5% (w/v) non-fat dried skim milk (NFM)for 1 h at room temperature. Membranes were then

incubated with bovine ALS antisera (1:4000 dilution) inblocking buffer overnight at 4 �C. To establish specificity,incubation with the primary antibody was performed inthe presence or absence of ALSC28. After incubation,membranes were washed in TBST and incubated withgoat anti-rabbit IgG-horseradish peroxidase (HRP)(1:5000 dilution) in blocking buffer for 1 h at roomtemperature. Signals were detected by LumiGLO westernblot chemiluminescence (KPL, Gaithersburg, MD, USA)and analyzed by densitometric scanning and NIH Image1·63 software.

Figure 1 Nucleotide and deduced amino acid sequence of bovine ALS. Amino acid sequence is shown below the nucleotide sequencein standard one-letter abbreviation. The nucleotide and amino acid sequence are numbered on the left relative to the translation initiationcodon (A+1TG). Amino acid 33 was identified by SignalP software analysis as the first amino acid residue of the mature protein and isboxed. The conserved 12 cysteine residues are indicated by bold and underlined letters. The stop codon TGA is indicated by a star.

Acid labile subunit in the dairy cow · J W KIM and others 585

www.endocrinology-journals.org Journal of Endocrinology (2006) 189, 583–593



Distribution of ALS in bovine fluids

Catheters were surgically implanted in the thoracic ductof six dairy calves using the procedures described byHartmann and Lascelles (1966). They were used to obtainmatched jugular blood and thoracic lymph samples. Bloodand follicular fluid were obtained from a second group offour cows (Porter et al. 2000). These cows were injectedwith a luteolytic dose of prostaglandin F2� (PGF2�)between days 7 and 14 of the estrous cycle. After the onsetof estrus, animals were examined by transrectal ultra-sonography to identify the dominant follicle, and thenovaries were removed by colpotomy. The dominantfollicle (�9 mm) was dissected and its follicular fluid wasaspirated using an 18-gauge needle. Finally, blood andcolostrum were obtained from six multiparous dairy cowsat day 245 of gestation (5 weeks before parturition).Plasma was prepared by heparin addition and centrifuga-tion (3000 g). Colostrum was centrifugated (10 000 g) for15 min and the aqueous phase was collected. All fluidswere frozen at �20 �C until analyzed for ALS abundanceby western immunoblotting (as described in previoussection).

Regulation of ALS during the transition from pregnancyto lactation

Temporal profiles of plasma ALS abundance and hepaticALS mRNA were studied in four multiparous cows in theperiod between 35 days prepartum and 56 days post-partum. Cows were fed a total mixed ration (TMR)according to recommended feeding standards (NRC2001); cows were fed once daily at 1100 h and wereallowed to feed ad libitum. The TMR contained 1·45 Mcalof net energy of lactation (NEL) and 125 g crude protein(CP) per kg of dry matter (DM) between days �35 and�21 (relative to parturition on day 0), 1·63 Mcal NEL/kgDM and 143 g CP/kg DM between day �21 andparturition, and 1·76 Mcal NEL/kg DM and 183 g CP/kgDM during lactation (NRC 2001). Cows were milkedthrice daily at 0900, 1600 and 2300 h. Blood samples werecollected between 0900 and 1000 h on days �35, �7,�1, +1, +3, +7, +10, +21 and +56 by coccygealvenipuncture. Plasma was prepared immediately andfrozen at �20 �C. After blood sampling on days �35,+3 and +56, liver biopsies were obtained by a non-invasive percutaneous method (Kim et al. 2004) andsnap-frozen in liquid N2.

To investigate the effect of GH on plasma ALS, asecond group of 12 cows was used in the period from5 weeks prepartum to 5 weeks postpartum. They were feda TMR ad libitum once daily according to feedingstandards (NRC 2001). The NEL and CP content of theTMR were 1·36 Mcal/kg DM and 130 g/kg DM duringweek �5, 1·56 Mcal/kg DM and 141 g/kg DM betweenweek �4 and parturition, and 1·72 Mcal/kg DM and179 g/kg DM after parturition. Cows were randomlyallocated to a control or a GH group. The GH groupreceived a daily injection of recombinant bovine GH(bovine somatotropin (bST), 45 mg i.m., lot 96J-B5128–002; Monsanto, St Louis, MO, USA) on three consecu-tive days during weeks �5, �2, +1 and +5. This periodof GH administration produces a maximal IGF-I responsethat persists for over 24 h after the last injection (Cohicket al. 1989). Injections were initiated on the second day ofeach week. Blood samples were obtained by coccygealvenipuncture between 1000 and 1130 h, the day beforethe first GH injection (day 0) and the day after the lastinjection (day 4). Control cows were sampled on the sameday of each week. Plasma was prepared immediately andfrozen at �20 �C.

Abundance of plasma ALS and hepatic ALS mRNAwas analyzed by western immunoblotting and RPA,respectively, as described above. Plasma IGF-I wasmeasured by a previously described double antibody RIA(Kim et al. 2004). Bovine recombinant IGF-I was used foriodination and standard (recombinant IGF-I, lot GST-3;Monsanto). Primary antibody against IGF-I (rabbit anti-human IGF-I, lot AFP4892898) was obtained from theNational Hormone and Pituitary Program (Bethesda,

Figure 2 Spatial ALS expression in bovine tissues. Top panel:tissues were obtained from two late lactation dairy cows. TotalRNA (20 �g) was analyzed simultaneously by RPA for theabundance of ALS mRNA and 18S rRNA. Bottom panel: theALS mRNA signal was normalized to the 18S signal and isrepresented as mean�S.E. (expressed as a percentage of theliver signal).





MD, USA). The secondary antibody was a caprine anti-rabbit IgG (lot 12515; Biotech Source, Inc, Franklin, MA,USA). Inter- and intra-assay coefficients of variationaveraged less than 8% and 9% respectively.

Statistical analysis

Levels of ALS in various fluids were expressed as apercentage of matched plasma levels and analyzed by amodel accounting for source (lymph, follicular fluid andcolostrum) as the fixed effects and animal as the randomeffect. Data obtained during the temporal study in transi-tion dairy cows were analyzed by a mixed model withtime as the fixed effect and animal as the random effect.When the effect of time was significant (P<0·05), thevariation was partitioned into linear, quadratic and cubiccontrasts. For the GH study, the response was calculatedfor each variable as the difference between day 4 and day0. Responses were analyzed by a model accounting forGH (control vs GH), time relative to parturition (weeks–5, �2, +1 and +5), and their interaction (GH�Time)as the fixed effects and animal as the random effect.

Results

Primary structure and spatial expression of bovine ALS

We isolated a partial 1833 bp bovine ALS cDNA contain-ing an open reading frame of 611 amino acids (Fig. 1).

The cDNA has a high degree of identity with its human(>75%) and ovine ALS counterparts (>95%). Computeranalysis of the amino acid sequence predicted a signalpeptide of 32 residues and a mature protein of 579residues. The deduced mature protein has all the featurespreviously described for ALS (Boisclair et al. 2001),including 12 cysteine residues clustered predominantly inthe amino and carboxyl terminal regions, 20 leucine-richrepeats of 24 residues and 6 potential asparagine-linkedglycosylation sites (NXS/T). Identities of mature bovineALS with chicken, mouse, human, pig and sheep ALSwere 60, 73, 76, 82 and 96% respectively.

We designed an RPA based on the bovine ALS cDNAand surveyed ALS expression in mature dairy cows. ALSexpression was approximately fivefold higher in liver thanin other tissues surveyed (Fig. 2). Expression in non-hepatic tissues was obvious in lung, small intestine, adiposetissue, kidney and heart, but nearly absent in brain andmuscle.

Detection of bovine ALS in plasma and extra-vascular fluid

To raise bovine ALS antisera, 3 rabbits were immunizedwith a peptide corresponding to the 28 carboxyl-terminalresidues of bovine ALS (ALSc28). This peptide is com-pletely conserved in sheep ALS but has less than 80%identity with the carboxyl termini of mouse, chicken, pigand human ALS (Fig. 3A). Antisera were tested bywestern immunoblotting for their ability to detect a

Figure 3 Development of bovine ALS antisera. (A) Cross-species comparison of the 28 carboxyl terminalresidues of ALS. For each species, amino acids differences from bovine residues are shaded. (B) Plasma(0·2 �l) from the indicated species was electrophoresed on a 10% reducing SDS–polyacrylamide gel. Afterelectroblotting, the membranes were incubated with bovine ALS antiserum 1082, 01 or 1085. Arrows onthe right correspond to the 84 kDa molecular mass marker. (C) Specificity of antiserum 1082. Bovineplasma (0·1–0·5 �l) was electrophoresed on a 10% reducing SDS–polyacrylamide gel. Immunoblotting wasperformed in the absence (�) or presence (+) of 10 �g/ml of the carboxyl terminal ALSC28 peptide.(D) The 84 kDa protein detected by antiserum 1082 is glycosylated. Bovine plasma (2 �l) was incubated inthe absence (�) or presence (+) of peptide-N-glycosidase F (PNGase F). Reactions were electrophoresedon a 10% reducing SDS–polyacrylamide gel. Immunoblotting was performed with pre-immune rabbit serum(PRE) or antiserum 1082.





�90 kDa protein in plasma. All three antisera detected aprotein of 84 kDa in cattle, sheep, goat and chickenplasma (Fig. 3B). Antiserum 1085 was also capable ofdetecting a �84 kDa signal in pig plasma, but none ofthese antisera detected ALS in mouse or human plasma.This cross-species reactivity profile suggests that theantisera recognized an epitope defined predominantly bythe 4 amino acid residues PPSL present in bovine, sheepand chicken ALS (Fig. 3A).

Additional studies were performed to validate anti-serum 1082. First, we assessed its specificity by determin-ing whether the 84 kDa signal present in bovine plasmawas eliminated when performing immunoblotting in thepresence of the ALSC28 peptide (Fig. 3C). As expected fora specific signal, the �84 kDa band was completelyeliminated in the presence of ALSC28. Secondly, wedetermined whether the 84 kDa protein is glycosylated.Bovine plasma was incubated in the presence or absenceof PNGase F, followed by SDS-PAGE and immuno-blotting (Fig. 3D). PNGase F incubation collapsed theimmune signal to 60 kDa, the size predicted by theprimary amino acid sequence of mature ALS. Overallthese studies indicate that ALS is the 84 kDa proteindetected in bovine plasma by antiserum 1082.

ALS abundance was compared in matched plasma andlymph samples obtained from growing cattle. As observed inhumans, ALS was more abundant in plasma than lymph(Fig. 4A). ALS has also been shown to occur in other fluidcompartments (Boisclair et al. 2001). To determine if thisoccurs in the cow, we sampled the mammary and ovarianfollicular fluid compartments (Fig. 4A). Mammary secre-tions were obtained from dairy cows 5 weeks before partu-rition and again after 5 weeks of lactation. ALS was presentin colostrum, but at only 15% of levels seen in matchedplasma. ALS was nearly undetectable in milk (not shown).ALS was also present in fluid obtained from the dominantfollicle at levels that were approximately 80% of those seenin matched plasma (Fig. 4A). These data show that ALSoccurs at significant levels outside the vascular compartment.

Given the presence of ALS in colostrum and follicularfluid, we performed ALS mRNA analysis on the associ-ated tissues. ALS mRNA was expressed in both mammarygland and ovary at levels similar to those seen in kidney(Fig. 4B). ALS mRNA was also present in isolatedmammary epithelial cells and complete ovarian follicles.These data suggest that mammary and follicular cellssynthesize a portion of the ALS found in associated fluidcompartments.

Figure 4 Relative abundance of ALS in biological fluids. (A) Left panel: plasma and the indicated fluidswere sampled from six dairy calves (lymph), six dairy cows 5 weeks before parturition (colostrum) or fourlactating dairy cows (follicular fluid). Indicated volumes were electrophoresed on a 10% reducingSDS–polyacrylamide gel and analyzed by immunoblotting with antiserum 1082. Right panel: bars representmeans�S.E. of ALS in each fluid relative to plasma. Bars with different letters differ at P<0·05.(B) Mammary glands (mammary tissue) and whole ovaries (ovarian tissue) were collected from lactatingdairy cows. Mammary epithelial cells were isolated from a single lactating mammary gland (mammary cell).Ovarian follicles were dissected from freshly excised cow ovaries and pooled (ovarian follicle). Total RNA(20 �g) was isolated from all tissues and cells and analyzed for ALS mRNA and 18S rRNA abundance.





Regulation of ALS during the periparturient period

We measured the variation in plasma IGF-I and ALS indairy cows between day 35 prepartum and 56 postpartum.The concentration of plasma IGF-I dropped by 50%during the first week postpartum, and remained depresseduntil the end of the study (Fig. 5A). Plasma ALS followedan identical pattern between day 35 prepartum and day 10postpartum (Fig. 5B). Unlike IGF-I, however, plasmaALS started to recover by day 21 postpartum and returnedto prepartum levels by day 56 postpartum (Fig. 5B,quadratic P<0·001). To determine whether this reflectedchanges in hepatic expression, ALS mRNA was measuredin liver biopsies collected on days �35, +3 and +56.ALS mRNA expression was significantly lower on day +3than on day �35, but recovered completely by day +56(Fig. 6, P<0·05). We conclude that hepatic ALS geneexpression accounts for the variation in plasma ALSduring the periparturient period.

Finally, we examined the ability of GH to increaseplasma ALS abundance at selected times before (5 and2 weeks prepartum) and after parturition (1 and 5 weekspostpartum). At each time, blood samples were collectedbefore (day 0) and the day after a 3-day period of GHtreatment (day 4). Control, untreated cows were sampledat identical times. GH significantly increased plasma ALSbefore parturition and 5 weeks postpartum, but this effectwas abolished during the first week postpartum (Fig. 7A,GH�Time P<0·05). Nearly identical effects of GH wereseen when the response considered was plasma IGF-I(Fig. 7B). We conclude that plasma ALS is positivelyregulated by GH but that this response is abrogated duringthe week following parturition.

Discussion

In mature animals, the circulating IGF system consists ofIGF-I and -II, IGFBP-1 to -6 and ALS (Baxter 1994, Ooi& Boisclair 1999). The effects of development, hormonesand diseases on tissue expression and plasma concen-trations have been studied extensively in rodents andhumans (Rajaram et al. 1997, Monzavi & Cohen 2002, LeRoith 2003). Similar efforts have taken place in cattle withthe notable exception of ALS (Cohick 1998, Breier 1999,Renaville et al. 2002). As a first step to address this lack ofinformation, we measured ALS expression across tissuesobtained from mature dairy cows. In agreement withsurveys performed in rodents, primates and other farmanimals (Boisclair et al. 2001), the highest level of ALSexpression was found in liver. Non-hepatic expression hasalso been suggested by recent reports. For example, lowlevels of ALS mRNA were detected in mouse kidney(Ueki et al. 2000) by northern blot analysis, and in the ratrenal cortex by in situ hybridization (Chin et al. 1994).Using an RPA, Lee et al. (2001) detected ALS expressionin pig muscle, spleen, ovary and uterus. In the present study,

we detected ALS mRNA in bovine heart, kidney, smallintestine, adipose tissue and lung with signals ranging from20 to 35% of that seen in liver. Excluding the liver, thehighest level of expression in our study occurred in lung,

Figure 5 Profile of plasma ALS abundance during theperiparturient period. Four multiparous dairy cows were studiedduring the period between day �35 and +56 (relative toparturition on day 0). (A) The plasma IGF-I concentration wasmeasured by RIA. Bars represent means�S.E. (n=4) of the plasmaIGF-I concentration. The effect of time was partitioned into linear(L), quadratic (Q) and cubic (C) contrasts. (B) Top panel: plasmaALS was measured by western immunoblotting. Data from tworepresentative cows are shown. Bottom panel: bars representmeans�S.E. (n=4) of the plasma ALS signal. The effect of time waspartitioned into linear (L), quadratic (Q) and cubic (C) contrasts.





a tissue not previously known to express ALS. We alsofound ALS mRNA in the mammary gland and ovarywhere expression was contributed, at least in part, by cellsspecific to these tissues (mammary epithelial cells andovarian follicular cells). In the case of the ovarian follicularcells, we did not determine whether both theca andgranulosa cells expressed ALS, but this seems likely basedon a recent study in the pig ovary (Wandji et al. 2000).Overall, our data indicate that the liver accounts for thebulk of circulating ALS, as it does in other species(Boisclair et al. 2001), but also raise the possibility thatother tissues contribute to extra-vascular ALS.

The relative abundance of ALS in plasma and otherfluid compartments has not been evaluated directly incattle, reflecting the lack of reactivity of antibodies raisedagainst human or rat ALS with their bovine counterpart(Baxter 1990, Baxter & Dai 1994). Using the COOH-terminal sequence of bovine ALS as an antigen, weobtained antisera detecting bovine, but not human ormouse ALS. As previously shown in humans (Baxter1990), we observed much higher ALS abundance inplasma than in lymph, consistent with the inability of ALSto cross the endothelial barrier efficiently (Binoux &Hossenlopp 1988). We also detected ALS in colostrumand ovarian follicular fluid. The high ALS abundance inthe latter is consistent with follicular cell synthesis, other-wise a steep gradient would be expected between plasmaand follicular fluid.

Irrespective of its origin, a more important issue iswhether or not extra-vascular ALS is functionally import-ant. Both IGF-I and IGF-II (IGFs) are thought to bepositive regulators of follicular growth whereas IGFBP-3is the most abundant IGFBP in follicular fluid (Fortuneet al. 2004, Webb et al. 2004). In cattle, total concentrationsof IGF-I and IGFBP-3 do not fluctuate significantly in

fluid obtained from dominant and subordinate follicles.Rather, IGF-I availability varies as a consequence ofchanges in the abundance of IGFBP-2, -4 and -5 (Fortuneet al. 2004, Webb et al. 2004). Thus, the dominant folliclehas lower levels of IGFBP-4 and -5 and higher levels offree IGF-I, whereas the opposite occurs in atretic follicles(Fortune et al. 2004, Webb et al. 2004). The presence ofhigh levels of ALS and IGFBP-3 along with significantlevels of IGFBP-5 suggests that IGFs mostly occur asternary complexes in bovine follicular fluid. This hasactually been demonstrated for IGF-I in human follicularfluid (Hughes et al. 1997). IGFs are also importantmammary growth factors (Marshman & Streuli 2002,Akers et al. 2005), including during the period of epithelialcell proliferation and differentiation of late pregnancy(Brisken et al. 2002) when we detected significant levels ofALS in colostrum. Late pregnancy is a period whenbovine colostrum contains high levels of IGF-I andIGFBP-3 (Baumrucker & Erondu 2000), raising thepossibility of significant ternary complex formation.Moreover, IGFBP-5 is expressed at its highest level duringlactogenesis in the bovine mammary gland (Plath-Gableret al. 2001). Additional studies are needed to determinewhether ternary complex formation occurs during thedifferent developmental stages of the bovine follicle andmammary gland.

Periparturient dairy cows provide a unique opportunityto examine the regulation of ALS during a physiologicalcondition that results in dynamic changes in plasma IGF-I.Dairy cows experience an abrupt, 60–70% decline inplasma IGF-I and IGFBP-3 between late pregnancy andparturition (Kobayashi et al. 1999, Kim et al. 2004,Boisclair et al. 2006). In agreement with others (Kobayashiet al. 1999, Radcliff et al. 2003), we showed that theplasma IGF-I depression is associated with reduced IGF-I

Figure 6 Hepatic ALS expression during the periparturient period. Four multiparous dairy cows werestudied during the period between day �35 and +56 (relative to parturition on day 0). Left panel: livertotal RNA (10 �g) was analyzed simultaneously for abundance of ALS mRNA and 18S RNA by RPA. Datafrom two representative cows are shown. Right panel: bars represent means�S.E. (n=4) of the ALS mRNAsignal normalized to the 18S signal. Bars with different letters differ at P<0·05.





mRNA in liver (Kim et al. 2004). We now show thatplasma ALS abundance is also decreased between preg-nancy and early lactation, an effect completely accountedfor by lower hepatic ALS mRNA levels. ALS has a lowaffinity for binary IGF:IGFBP-3 complexes and mustcirculate in excess over IGF-I and IGFBP-3 to efficientlyrecruit these proteins into long-lived ternary complexes(Holman & Baxter 1996, Boisclair et al. 2001). Therefore,reduced ALS synthesis in the days following parturitionprobably amplifies the effect of decreased hepatic IGF-Iproduction on plasma IGF-I. Within 3 weeks of parturi-tion, however, plasma ALS levels started to recover suchthat both hepatic ALS expression and plasma ALS hadreturned to pre-calving levels by day 56 of lactation.Hepatic IGF-I mRNA abundance was also completelynormalized in these cows by day 56 (Kim et al. 2004), butthe plasma IGF-I concentration was not. These resultssuggest that ALS plays little or no role in the persistingplasma IGF-I depression.

In rodents and humans, GH is the most potent positiveregulator of hepatic ALS mRNA abundance and plasmaALS concentration (Baxter & Dai 1994, Dai & Baxter1994, Olivecrona et al. 1999). The essential role of GH isillustrated by the absence of ALS and ternary complexes inthe plasma of GH-deficient animals (Zapf et al. 1995). Wepreviously showed that these effects of GH are transcrip-tional and mediated by STAT5 binding to a single GASelement in the proximal promoter of the mouse ALS gene(Ooi et al. 1998). This GAS element is completelyconserved in the promoter of the rat, sheep and humanALS gene (Delhanty & Baxter 1997, Rhoads et al. 2000,Suwanichkul et al. 2000). A similar mechanism hasrecently been shown to explain the positive effects of GHon the transcription of the rat IGF-I gene (Woelfle et al.2003, Chia et al. 2006). In the present study, GHtreatment increased levels of both ALS and IGF-I duringpregnancy and week 5 of lactation, but was ineffective1 week after parturition. In this context, we recentlydemonstrated that the liver-specific GH receptor tran-script and hepatic GH receptor abundance were decreasedimmediately after parturition but returned to normalwithin 3 weeks of parturition (Kim et al. 2004). There-fore, loss of hepatic GH action probably accounts fordecreased production of both ALS and IGF-I in theimmediate postparturient period.

Dairy cattle offer other opportunities to study the regu-lation of ALS, including other conditions where plasmaIGF-I increases (euglycemic–hyperinsulinemic clamp) ordecreases rapidly (50–80% decline during under-nutrition) (McGuire et al. 1995a, 1995b). Cattle are alsorecognized as a valuable model for studying mammary andovarian follicular development (Fortune et al. 2004, Akerset al. 2005) and have easily accessible fluid compartmentsassociated with these tissues. The reagents we have devel-oped will facilitate studies on the role of ALS and ternarycomplexes in plasma and other fluid compartments of cattle.

Figure 7 Effect of GH on plasma ALS during the periparturientperiod. Twelve multiparous cows were studied in the periodbetween –5 and +5 weeks (relative to parturition on day 0).Cows were randomly allocated to remain untreated (control, n=6)or to receive three consecutive daily injections of recombinant GH(45 mg/injection, n=6) during each of weeks �5, �2, +1 and+5. During these weeks, blood samples were obtainedimmediately before (day 0) and the day following the 3-day periodof GH injection (day 4). Blood samples were obtained at similartimes from the control group. (A) Top panel: plasma ALS wasmeasured for each sample by western immunoblotting. Data fromtwo representative cows are shown. Bottom panel: the responsein plasma ALS (� plasma ALS) was calculated as the plasmaabundance difference between day 4 and day 0. Bars representmeans�S.E. (n=6) of the response in plasma ALS. (B) Plasma IGF-Iwas measured by RIA. The response in plasma IGF-I (� plasmaIGF-I) was calculated as the plasma concentration differencebetween day 4 and day 0. Bars represent means�S.E. (n=6) ofthe plasma IGF-I response.





Acknowledgements

The authors thank D A Porter and S M Quirk forproviding the ovarian follicular fluid.

Funding

This work was supported by the Cornell UniversityAgricultural Experiment Station. J W K was supported inpart by Korea Government Study-Abroad Scholarship.The authors declare that there is no conflict of interest thatwould prejudice the impartiality of this scientific work.

References

Akers RM, Ellis SE & Berry SD 2005 Ovarian and IGF-I axis controlof mammary development in prepubertal heifers. Domestic AnimalEndocrinology 29 259–267.

Baumrucker CR & Erondu NE 2000 Insulin-like growth factor (IGF)system in the bovine mammary gland and milk. Journal of MammaryGland Biology and Neoplasia 5 53–64.

Baxter RC 1990 Circulating levels and molecular distribution of theacid-labile (alpha) subunit of the high molecular weight insulin-likegrowth factor-binding protein complex. Journal of ClinicalEndocrinology and Metabolism 70 1347–1353.

Baxter RC 1994 Insulin-like growth factor binding proteins in thehuman circulation: a review. Hormone Research 42 140–144.

Baxter RC & Dai J 1994 Purification and characterization of theacid-labile subunit of rat serum insulin-like growth factor bindingprotein complex. Endocrinology 134 848–852.

Bell AW 1995 Regulation of organic nutrient metabolism duringtransition from late pregnancy to early lactation. Journal of AnimalScience 73 2804–2819.

Binoux M & Hossenlopp P 1988 Insulin-like growth factor (IGF) andIGF-binding proteins: comparison of human serum and lymph.Journal of Clinical Endocrinology and Metabolism 67 509–514.

Boisclair YR, Rhoads RP, Ueki I, Wang J & Ooi GT 2001 Theacid-labile subunit (ALS) of the 150 kDa IGF-binding proteincomplex: an important but forgotten component of the circulatingIGF system. Journal of Endocrinology 170 63–70.

Boisclair YB, Wesolowski SR, Kim JW & Ehrhardt RA 2006 Rolesof growth hormone and leptin in the periparturient dairy cow. InRuminant Physiology: Digestion, Metabolism and Impact of Nutrition onGene Expression, Immunology and Stress, pp 327–346. Eds K Sejrsen,T Hvelplund & MO Nielsen. Wageningen, The Netherlands:Wageningen Academic Publishers.

Breier BH 1999 Regulation of protein and energy metabolism by thesomatotropic axis. Domestic Animal Endocrinology 17 209–218.

Brisken C, Ayyannan A, Nguyen C, Heineman A, Reinhardt F, Tan T,Dey SK, Dotto GP & Weinberg RA 2002 IGF-2 is a mediator ofprolactin-induced morphogenesis in the breast. Developmental Cell3 877–887.

Chia DJ, Ono M, Woelfle J, Schlessinger-Massart M, Jiang H &Rotwein P 2006 Characterization of distinct stat5b binding sitesthat mediate growth hormone-stimulated IGF-I gene transcription.Journal of Biological Chemistry 281 3190–3197.

Chin E, Zhou J, Dai J, Baxter RC & Bondy CA 1994 Cellularlocalization and regulation of gene expression for components ofthe insulin-like growth factor ternary binding protein complex.Endocrinology 134 2498–2504.

Cohick WS 1998 Role of the insulin-like growth factors and theirbinding proteins in lactation. Journal of Dairy Science 81 1769–1777.

Cohick WS, Plaut K, Sechen SJ & Bauman DE 1989 Temporalpattern of insulin-like growth factor-I response to exogenousbovine somatotropin in lactating cows. Domestic AnimalEndocrinology 6 263–273.

Dai J & Baxter RC 1994 Regulation in vivo of the acid-labile subunitof the rat serum insulin-like growth factor-binding proteincomplex. Endocrinology 135 2335–2341.

Delhanty PJ & Baxter RC 1997 Cloning and characterization of therat gene for the acid-labile subunit of the insulin-like growth factorbinding protein complex. Journal of Molecular Endocrinology19 267–277.

Domene HM, Bengolea MS, Martinez AS, Ropelato MG, Pennisi P,Scaglia P, Heinrich JJ & Jasper HG 2004 Deficiency of thecirculating insulin-like growth factor system associated withinactivation of the acid-labile subunit gene. New England Journal ofMedicine 350 570–577.

Fortune JE, Rivera GM & Yang MY 2004 Follicular development:the role of the follicular microenvironment in selection of thedominant follicle. Animal Reproduction Science 82–83 109–126.

Guler HP, Zapf J, Schmid C & Froesch ER 1989 Insulin-likegrowth factors I and II in healthy man. Estimations of half-livesand production rates. Acta Endocrinologica (Copenhagen)121 753–758.

Hartmann PE & Lascelles AK 1966 The flow and lipoid compositionof thoracic duct lymph in the grazing cow. Journal of Physiology184 193–202.

Holman SR & Baxter RC 1996 Insulin-like growth factor bindingprotein-3: factors affecting binary and ternary complex formation.Growth Regulation 6 42–47.

Hughes SC, Mason HD, Franks S & Holly JM 1997 The insulin-likegrowth factors (IGFs) in follicular fluid are predominantly bound inthe ternary complex. Journal of Endocrinology 155 R1–4.

Kim JW, Rhoads RP, Block SS, Overton TR, Frank SJ & BoisclairYR 2004 Dairy cows experience selective reduction of the hepaticgrowth hormone receptor during the periparturient period. Journalof Endocrinology 181 281–290.

Kobayashi Y, Boyd CK, Bracken CJ, Lamberson WR, Keisler DH &Lucy MC 1999 Reduced growth hormone receptor (GHR)messenger ribonucleic acid in liver of periparturient cattle is causedby a specific down-regulation of GHR 1A that is associated withdecreased insulin-like growth factor I. Endocrinology 140 3947–3954.

Le Roith D 2003 The insulin-like growth factor system. ExperimentalDiabesity Research 4 205–212.

Lee CY, Kwak I, Chung CS, Choi WS, Simmen RC & Simmen FA2001 Molecular cloning of the porcine acid-labile subunit (ALS) ofthe insulin-like growth factor-binding protein complex anddetection of ALS gene expression in hepatic and non-hepatictissues. Journal of Molecular Endocrinology 26 135–144.

Marshman E & Streuli CH 2002 Insulin-like growth factors andinsulin-like growth factor binding proteins in mammary glandfunction. Breast Cancer Research 4 231–239.

Matitashvili E & Bauman DE 1999 Culture of primary bovinemammary epithelial cells. In Vitro Cellular and DevelopmentalBiology: Animal 35 431–434.

McGuire MA, Bauman DE, Dwyer DA & Cohick WS 1995aNutritional modulation of the somatotropin/insulin-like growthfactor system: response to feed deprivation in lactating cows. Journalof Nutrition 125 493–502.

McGuire MA, Dwyer DA, Harrell RJ & Bauman DE 1995b Insulinregulates circulating insulin-like growth factors and some of theirbinding proteins in lactating cows. American Journal of Physiology269 E723–730.

Monzavi R & Cohen P 2002 IGFs and IGFBPs: role in health anddisease. Best Practice and Research. Clinical Endocrinology andMetababolism 16 433–447.

NRC 2001 Nutrient Requirements of Dairy Cattle, 7th revised edn.Washington, DC, USA: National Academy Press.





Olivecrona H, Hilding A, Ekstrom C, Barle H, Nyberg B, Moller C,Delhanty PJ, Baxter RC, Angelin B, Ekstrom TJ et al. 1999 Acuteand short-term effects of growth hormone on insulin-like growthfactors and their binding proteins: serum levels and hepaticmessenger ribonucleic acid responses in humans. Journal of ClinicalEndocrinology and Metabolism 84 553–560.

Ooi GT & Boisclair YB 1999 Molecular biology of the insulin-likegrowth factor binding proteins. In The IGF System: MolecularBiology, Physiology and Clinical Applications, pp 111–139. Eds RGRosenfeld & CT Roberts Jr. New Jersey, USA: Humana Press.

Ooi GT, Cohen FJ, Tseng LY, Rechler MM & Boisclair YR 1997Growth hormone stimulates transcription of the gene encoding theacid-labile subunit (ALS) of the circulating insulin-like growthfactor-binding protein complex and ALS promoter activity in ratliver. Molecular Endocrinology 11 997–1007.

Ooi GT, Hurst KR, Poy MN, Rechler MM & Boisclair YR 1998Binding of STAT5a and STAT5b to a single element resembling agamma-interferon-activated sequence mediates the growthhormone induction of the mouse acid-labile subunit promoter inliver cells. Molecular Endocrinology 12 675–687.

Plath-Gabler A, Gabler C, Sinowatz, F, Berisha B & Schams D 2001The expression of the IGF family and GH receptor in the bovinemammary gland. Journal of Endocrinology 168 39–48.

Porter DA, Vickers SL, Cowan RG, Huber SC & Quirk SM 2000Expression and function of Fas antigen vary in bovine granulosaand theca cells during ovarian follicular development and atresia.Biology of Reproduction 62 62–66.

Radcliff RP, McCormack BL, Crooker BA & Lucy MC 2003 Plasmahormones and expression of growth hormone receptor andinsulin-like growth factor-I mRNA in hepatic tissue ofperiparturient dairy cows. Journal of Dairy Science 86 3920–3926.

Rajaram S, Baylink DJ & Mohan S 1997 Insulin-like growthfactor-binding proteins in serum and other biological fluids:regulation and functions. Endocrine Review 18 801–831.

Renaville R, Hammadi M & Portetelle D 2002 Role of thesomatotropic axis in the mammalian metabolism. Domestic AnimalEndocrinology 23 351–360.

Rhoads RP, Greenwood PL, Bell AW & Boisclair YR 2000Organization and regulation of the gene encoding the sheepacid-labile subunit of the 150-kilodalton insulin-like growthfactor-binding protein complex. Endocrinology 141 1425–1433.

Suwanichkul A, Boisclair YR, Olney RC, Durham SK & Powell DR2000 Conservation of a growth hormone-responsive promoter

element in the human and mouse acid-labile subunit genes.Endocrinology 141 833–838.

Twigg SM & Baxter RC 1998 Insulin-like growth factor(IGF)-binding protein 5 forms an alternative ternary complex withIGFs and the acid-labile subunit. Journal of Biological Chemistry273 6074–6079.

Ueki I, Ooi GT, Tremblay ML, Hurst KR, Bach LA & Boisclair YR2000 Inactivation of the acid labile subunit gene in mice results inmild retardation of postnatal growth despite profound disruptionsin the circulating insulin-like growth factor system. PNAS97 6868–6873.

Wandji SA, Gadsby JE, Simmen FA, Barber JA & Hammond JM2000 Porcine ovarian cells express messenger ribonucleic acids forthe acid-labile subunit and insulin-like growth factor bindingprotein-3 during follicular and luteal phases of the estrous cycle.Endocrinology 141 2638–2647.

Webb R, Garnsworthy PC, Gong JG & Armstrong DG 2004 Controlof follicular growth: local interactions and nutritional influences.Journal of Animal Science 82 (E-Suppl) E63–74.

Woelfle J, Chia DJ & Rotwein P 2003 Mechanisms of growthhormone action: identification of conserved Stat5 binding sites thatmediate GH-induced insulin-like growth factor-I gene activation.Journal of Biological Chemistry 278 51261–51266.

Yakar S, Rosen CJ, Beamer WG, Ackert-Bicknell CL, Wu Y, Liu JL,Ooi GT, Setser J, Frystyk J, Boisclair YR et al. 2002 Circulatinglevels of IGF-1 directly regulate bone growth and density. Journal ofClinical Investigation 110 771–781.

Yakar S, Boisclair YB, Pintar JE, Rosen C, Cohen P, Sun H,Stannard B & LeRoith D 2005 Triple gene inactivation of liverigf-I, the acid labile subunit and the IGFBP-3 leads to undectectablelevels of serum IGF-I. Proceedings of The Endocrine Society,87th Annual Meeting, San Diego, CA, USA. OR12–11, p 88.

Zapf J, Hauri C, Futo E, Hussain M, Rutishauser J, Maack CA &Froesch ER 1995 Intravenously injected insulin-like growth factor(IGF) I/IGF binding protein-3 complex exerts insulin-like effectsin hypophysectomized, but not in normal rats. Journal of ClinicalInvestigation 95 179–186.

Received in final form 14 March 2006Accepted 21 March 2006Made available online as an Accepted Preprint24 March 2006





Documents

Isolation of the cDNA encoding the acid labile subunit