Wasting in the acquired immune deficiency syndrome is associated with multiple defects in the serum...

Clinical Endocrinology (1996) 44, 501–514

Wasting in the acquired immune deficiency syndromeis associated with multiple defects in the seruminsulin-like growth factor system

Robert A. Frost*, Jack Fuhrers ˇ , Roy Steigbigelsˇ ,Peter Mariuzs , Charles H. LangU and Marie C.Gelato*Department of Medicine, *Divisions of Endocrinology andsInfectious Diseases and theUDepartment of Surgery, State University of New Yorkat Stony Brook, Stotny Brook, NY 11794

(Received 7 June 1995; returned for revision 27 July 1995;finally revised 28 September 1995; accepted 6 February 1996)

Summary

OBJECTIVE The aim of this investigation was to char-acterize the GH–IGF axis of patients with AIDS asso-ciated wasting. A special emphasis was placed ondetermining whether IGF binding proteins (IGFBPs) ofpatients who have lost more than 10% of their idealbody mass are structurally different from the IGFBPsof patients with no weight loss.DESIGN AND PATIENTS A cross-sectional study of 11AIDS patients was performed to determine whetherthe IGF system is abnormal in AIDS patients withwasting. Seven additional AIDS patients were fol-lowed longitudinally to determine whether AIDSpatients experience long-term changes to their IGFsystem.MEASUREMENTS Serum levels of GH and IGF-I weremeasured by radioimmunoassay, IGF-II was mea-sured by radioreceptor assay, and IGFBP-1 was mea-sured by an enzyme linked immunoassay. IGFBP-3and IGFBP-3 protease activity were measured byligand blotting and a BP-3 protease assay, respec-tively. IGFBP-3 ternary complex formation and IGFBP-1 phosphovariants were analysed by non-denaturingPAGE.RESULTS AIDS patients who had lost more than 10%of their ideal body mass demonstrated a 55% reduc-tion in serum IGF-I (81 vs 179¹g/l) and a 70% reduc-tion in IGF-II (226 vs 776¹g/l), compared to healthy

HIV negative subjects. IGF-I levels were depressed, insome patients, despite high serum levels of GH. AIDSpatients who had lost more than 10% of their idealbody mass had low levels of IGFBP-3 and a reducedability to form the IGFBP-3 ternary complex. TheIGFBP-3 ternary complex could be restored onlyupon addition of pure IGFBP-3 and acid labile subunitto serum. Serum IGFBP-1 was increased more thanthreefold compared to control subjects (90 vs 24¹g/l).IGFBP-1 was present as a free phosphoprotein inAIDS patients with low levels of IGF-I and in a boundform when serum IGF-I levels were normal. Changesin the GH–IGF axis were sustained for up to 25 monthsin AIDS patients with wasting.CONCLUSIONS AIDS wasting is associated with a GHresistant state that results in low levels of serum IGF-I,IGF-II and IGFBP-3, elevated levels of phosphorylatedIGFBP-1, and a reduced ability to form the IGFBP-3ternary complex.

Weight loss is a common manifestation of patients with theacquired immunodeficiency syndrome (AIDS) (Grunfeld &Feingold, 1992)1,2. The wasting syndrome accompanying thisdisease is multifactorial and, in part, due to inadequate foodintake and malabsorption, especially in the later stages of thedisease (Baumet al., 1994). However, a reduction in lean bodymass (LBM) has also been demonstrated in this patientpopulation prior to changes in caloric intake, suggesting ahypermetabolic catabolic state (Grinspoon & Bilezikian, 1992;Stein et al., 1990). The mechanism leading to an erosion ofLBM following infection with human immunodeficiency virus(HIV) and the development of AIDS has not been elucidated.

The GH–insulin-like growth factor (IGF) axis is thought tobe an essential component of the endocrine system responsiblefor stimulating protein synthesis and maintaining LBM inadults (Press, 1988; Salomonet al., 1991). Both GH and IGF-I,enhance nitrogen retention and influence muscle proteinsynthesis (Clemmonset al., 1987; 1992; Kupferet al., 1993;Fryburg et al., 1991). At the cellular level, IGF-I inhibitsprotein breakdown and stimulates amino acid uptake (Roederet al., 1986; Shimizuet al., 1986).

IGF physiology is greatly influenced by IGF binding proteins(IGFBPs). Six structurally homologous IGFBPs have beenidentified (Shimasaki & Ling, 1992). The IGFBPs are the

501# 1996 Blackwell Science Ltd

Correspondence: Dr Robert A. Frost, Department of Medicine,Division of Endocrinology, Health Sciences Center, T15, Room 060,SUNY at Stony Brook, Stony Brook, NY 11794, USA. Fax: 516 4443410.

products of six different genes and exist in multiple forms dueto their post-translational modification (Cohick & Clemmons,1993). More than 90% of IGF-I in serum is bound by IGFBP-3as part of a high molecular weight complex. IGF-I has itsgreatest anabolic effects when present as part of this complex(Kupferet al., 1993). Since changes in the concentration and/orstructure of IGFBPs dramatically influence the biologicalactivity of the IGFs (Clemmonset al., 1991), the quantitationof serum IGF-I alone may be an insufficient measure of thepotential biological activity of the IGFs. IGF-I also affects theconcentration and structure of IGF binding proteins (Baxteretal., 1993; Conoveret al., 1993). The influence of the IGFBPs onIGFs and vice versa supports the contention that the IGF systemshould be considered as a whole when evaluating IGFphysiology (Langford & Miell, 1993).

Abnormalities in the IGF system have been found to beassociated with various catabolic states (Dahnet al., 1988;Frostet al., 1994; Rosset al., 1991; Wojnaret al., 1995), suchas is seen in terminally ill AIDS patients. An understanding ofthe IGF system is important in AIDS because of the knownanabolic effects of GH and IGF-I. Both peptides have beenproposed as pharmacological agents for treating AIDS associatedwasting (Mulliganet al., 1993; Liebermanet al., 1994). Thepurpose of the present study was to determine whether changesoccur in the circulating levels of GH, IGF-I and IGF-II in AIDSpatients with wasting and whether alterations in IGFBPs could bepartly responsible for these changes.

Subjects and methods

Subjects

This study was approved by the committee on research

involving human subjects at SUNY-Stony Brook. All patientsgave informed consent. Blood samples were obtained byvenipuncture from 11 HIV positive, non-fasted adults, and 10normal healthy volunteers after an 18 hour fast. Serum sampleswere allowed to clot at room temperature for 15–30 minutesand were then centrifuged at 600g for 20 minutes. Serum wasimmediately decanted and stored at –208C. Patient character-istics are described in Table 1. Serum from seven AIDS patientswho were studied over a 25-month period were examinedretrospectively.

Acid–ethanol extraction of serum

Fifty microliters of serum was mixed with 200�l of acidifiedethanol (0.25M HCl/87.5% ethanol) and incubated at roomtemperature for 30 minutes. The resulting precipitate (contain-ing IGFBPs) was centrifuged at 10 000g for 10 minutes. A 125-�l aliquot of the supernatant (containing IGFs) was neutralizedwith 50�l of 0.855 M Tris-base. Neutralized samples wereincubated at –208C overnight. The presence of ethanol in thesample prevents the sample from freezing. An additionalcentrifugation for 10 minutes at 10 000g was performed. Analiquot of the resulting supernatant was diluted in either IGF-Iradioimmunoassay (RIA) or IGF-II radioreceptor assay (RRA)buffer and quantitated as described below. Acid–ethanolextraction removed more than 99% of the IGF-binding activityand IGFBP-1 from serum.

IGF-I RIA

IGF-I was measured by RIA using an antibody generatedagainst human IGF-I and supplied by the National PituitaryProgram as described previously (Olchovskyet al., 1991).

502 R.A. Frost et al.

# 1996 Blackwell Science Ltd.Clinical Endocrinology, 44, 501–514

Patient Weight loss Serum albumin CD4 (�) Opportunisticno. Sex (% IBM) (g/L) (cells/mm3) infection*

1 M <10 45 10 PCP ��

2 M <10 29 150 TB (ÿ)3 F 26 14 3 MAI (�)4 F 19 23 5 MAI �ÿ�

5 M 32 15 10 PCP �ÿ�

6 M 15 25 13 PCP ��

7 F 25 26 40 PCP ��

8 F <10 32 20 PCP �ÿ�

9 M 15 33 200 PA, PCP �ÿ�

10 M <10 31 0 PCP, KS ��

11 F 20 10 10 MAI ��

1PCP,Pneumocystis cariniipneumonia; TB,Mycobacteriumtuberculosis; MAI,Mycobacteriumavium intracellular complex; PA,Pseudomonas aeruginosa; KS, Kaposi’s sarcoma.� andÿdenote the presence of an active infection for which the patient was on therapy at the time bloodwas collected).

Table 1 Clinical characteristics of patientswith AIDS

Briefly, each assay tube contained IGF-I antibody,125I-IGF-I (10000 c.p.m./tube), and IGF-I standard (Upstate BiotechnologyInc., Lake Placid, NY) or unknown sample in a total volume of0.2 ml assay buffer. Bound and free radioactivity wereseparated using Pansorbin cells after an overnight incubation.The ED50 for the assay was 0.04–0.08 ng/tube; intra-assay andinterassay coefficients of variation were 7 and 10%, respec-tively. Addition of an excess of IGFBP-1 (1500�g/l) to samplesassayed by this method had no effect on IGF-I measurementswhen protamine sulphate was included in the assay buffer.

IGF-II Radioreceptor assay

IGF-II was measured by a RRA using Triton X-100 solubilizedIGF-II/mannose-6-phosphate receptors from rat placentalmembranes, as described previously (Gelatoet al., 1992).Briefly, a volume of solubilized IGF-II receptor that provided75% maximal binding of 20 000 c.p.m./tube of125I-IGF-II wasused. IGF-II standard (Upstate Biotechnology Inc., Lake Placid,NY) or unknown sample was assayed in a total volume of 0.4 mlassay buffer (Dulbecco’s modified Eagle’s medium with 0.5%BSA, Sigma Chemical, St Louis, MO). Standards and acid–ethanol extracts of serum were run in duplicate. After anovernight incubation at 48C, bound and free tracer wereseparated using a 5% solution of BSA-activated charcoal.Radioactivity in an aliquot of the resulting supernatant wasmeasured. The ED50 for the assay was 3–4�g/l; the intra andinter-assay coefficients of variation were 10 and 15%,respectively.

Growth hormone, insulin, and cortisol RIA

Growth hormone, cortisol, and insulin were measured in serumwith commercially available RIA kits from Diagnostic ProductsCorporation (Los Angeles, CA).

IGF/IGFBP complex formation and non-denaturingPAGE

125I-GF-I or -II (1 � 105 c.p.m.) was incubated with 6�l ofpatient sera in 0.4 M NaH2PO4 (pH 7.4) for 2 hours at roomtemperature. The resulting125I-IGF/IGFBP complexes wereresolved on non-denaturing PAGE gels consisting of stacking(4% acrylamide) and separating (7.5% acrylamide) compo-nents. Some serum samples were incubated with IGF-I (250 ng),IGFBP-1 (Medix Biochemica, Kaunianen, Finland), IGFBP-3(UBI, Lake Placid, NY) or acid labile subunit (ALS) antibody(Dr Robert Baxter, New South Wales, Australia) for 1.5 hours.Gel electrophoresis was performed at 38 mA constant currentfor 4.5 hours. The resulting gels were washed with 0.2%glycerol and dried under vacuum. The dried gel was usedto expose X-ray film (Fuji RX film, Japan) at –708C for

1.5–15 hours. Under the electrophoresis conditions describedhere, 125I-IGF-I/IGFBP-1 complexes migrate as 5 bands inthe bottom third of the gel (Frost & Tseng, 1991).125I-IGF-II/IGFBP-3 complexes migrate as a single broad band in theupper third of the gel.125I-IGFBP-3 incubated with serumforms a complex that co-migrates with the ternary complexisolated from the 150 kDa peak off a Sephadex G-200column. The ternary complex can also be formed by mixing125I-IGF-II with pure ALS (Diagnostic Systems Laboratories,Webster TX) and IGFBP-3, but not with either componentalone.

IGFBP-3 proteolysis

IGFBP-3 proteolysis was measured by incubating125I- IGFBP-3 (30 000 c.p.m., Diagnostic Systems Laboratories, Wester,TX) with 4 �l of serum in Tris buffer for 6 hours at 378C asdescribed previously (Bereketet al., 1995). The reaction wasstopped by boiling the sample in sample buffer. The reactionproducts were separated by SDS-PAGE.125I-IGFBP-3 proteo-lysis in control serum was compared to proteolysis by serumfrom 11 AIDS patients. The proteolysis reaction products werealso run on a non-denaturing PAGE gel to examine ternarycomplex formation. These samples were not boiled.

Western blotting

Serum samples were electrophoresed on either denaturing ornon-denaturing PAGE gels as described above. Proteins wereelectrophoretically transferred to nitrocellulose with a semi-dryblotter (Bio-Rad Laboratories, Melville, NY) set at 21 Vconstant voltage for 15–20 minutes. The resulting blot wasblocked with 1% BSA for 1.5 hours and incubated withantibodies against either IGFBP-1 (clone 6303, MedixBiochemica, Kaunianen, Finland) or IGFBP-3 (Upstate Bio-technology, Lake Placid, NY) at a 1 : 1000 dilution. Excessprimary antibody was washed away with Tris-buffered salinecontaining 0.5% Tween-20 (TBS-Tween). Blots were incu-bated with either anti-mouse or anti-rabbit immunoglobulinconjugated with horse-radish peroxidase for 2 hours (Sigma, St.Louis, MO). Blots were washed with TBS-Tween and brieflyincubated with the components of an enhanced chemilumines-cent detection system (Amersham, Arlington Hts, IL). Thedried blot was used to expose X-ray film for 1–3 minutes.

IGFBP-1 assay

An IGFBP-1 ELISA was developed using IGFBP-1 mono-clonal antibodies. Monoclonal antibody 6301 (Medix Bio-chemica, Kaunianen Finland) was used as a capture antibody.Monoclonal antibody 6303 was conjugated to horse-radishperoxidase and used as a detection antibody. Antibody 6301

IGFs and IGFBPs in AIDS wasting 503

# 1996 Blackwell Science Ltd.Clinical Endocrinology,44, 501–514

was coated on Maxisorb microtitre plates (Nunc, Naperville,IL) at a 1:1000 dilution. The plate was blocked with 1% BSA inTBS for 1–2 hours. Standards and samples were diluted in TBScontaining 5% fetal bovine serum (Hyclone, Logan, UT) andincubated with the coated plate overnight at 48C. The plate waswashed with TBS-Tween and incubated with the 6303-horse-radish peroxidase conjugate at a 1 : 1000 dilution for 4 hours.IGFBP-1 was detected by the addition of 0.4 g/l of o-phenylenediamine dissolved in 24 mM citric acid, 51 mM sodium phos-phate and 0.012% hydrogen peroxide at pH 5. The reaction wasstopped with 2.25M sulphuric acid. The plate was read on amicrotitre plate reader (Bio-tek Instruments, Rutherfod NJ) at awavelength of 492 nm. IGFBP-1 was purified from humanendometrial stromal cell conditioned medium as describedpreviously and used as standard (Frost &Tseng, 1991). TheIGFBP-1 monoclonal antibodies used in this assay recognizedall the phosphovariants of IGFBP-1. IGFBP-1 concentrationsobtained by this assay were equivalent to those obtainedusing a commercially available IGFBP-1 immunoradiometricassay (Diagnostic Systems Laboratory, TX; unpublishedobservation).

Results

Patient population

Of the 11 AIDS patients examined, seven had experienced aweight loss of more than 10% of their ideal body mass (Table 1).All the patients had AIDS defining opportunistic infections andCD4 + lymphocyte counts of 200 cells/mm3 or below. At thetime blood was collected, six of the patients had infections forwhich they were being treated (Table 1). Weight loss was notassociated with the presence of acute infections. Patientswho had lost more than 10% of their IBM demonstrated a36% reduction in serum albumin, were severely insulinopaenic(–88%), and had elevated levels of serum cortisol (+94%)compared to control subjects (Table 2).

Serum GH levels

Although GH secretion is episodic, very high levels of serumGH are often indicative of abnormalities in the GH/IGF-I axis(Laronet al., 1968). We measured GH in the 11 AIDS patients.Three of the AIDS patients showed serum levels of GH nearly50-fold greater than control values (Fig. 1a, patients 3,4 and11). These same patients had very low levels of IGF-I (73�g/l)(Fig. 1a).

Serum IGF-I and IGF-II levels

In the 11 AIDS patients examined there was considerablevariation in the levels of serum IGF-I and -II (Fig. 1b and c).



Fig. 1 a, Serum GH levels were assayed in the 11 randomly selectedAIDS patients, described in Table 1, by RIA. Results are expressedas�g/l GH (1�g=lGH � 2 mU/l). In b and c, serum IGF-I and IGF-IIwere assayed in the same patients by an IGF-I RIA and IGF-IIradioreceptor assay, respectively. Results are expressed as�g/l ofIGF-I and IGF-II. Patient identification is the same as that in Table 1.

Table 2 Hormone levels in AIDS patients with greater than a 10% lossof IBM

Controls AIDS Patients

IGF-I (�g/l) 179� 21 81� 20*IGF-II (�g/l) 776� 220 226� 40*Insulin (pmol/l) 237� 22 29� 29*Cortisol (nmol/l) 331� 28 662� 331IGFBP-1 (�g/l) 24� 15 90� 43Albumin (g/l) 42� 6 26� 8

* Values are means�SE,n= 7, P< 0.05.

Despite this variation, AIDS patients who had lost more than10% of their ideal body mass had lower circulating concentra-tions of IGF-I (–55%) and IGF-II (–70%) than did healthycontrol subjects (Fig. 1b and c, Table 2). Reduced serum IGFlevels were not associated with the presence of an acuteinfection. IGF-I-values measured by RIA were not affected byIGFBP-1 at a concentration tenfold higher than that found inAIDS sera (Fig. 2). Acid–ethanol extraction removed more than99% of the IGF-binding activity and IGFBP-1 from serum(Fig. 2, inset, compare lanes 1 and 2).

Serum IGFBP-3 levels

Low serum concentrations of IGF-I and IGF-II in AIDS patientssuggested that a defect may exist in the ability of IGFBP-3 tosequester these peptides. IGFBP-3 in serum from AIDS patientswas compared with control sera by ligand blotting (Fig. 3).When125I-IGF-I and II were used as the radioligand, IGFBP-3was detected as a 39/41 kDa doublet in both control and AIDSsera. Those patients who had lost more than 10% of their idealbody mass had less IGFBP-3 than age-matched controls

(Patients 2–5 and 7–11, Fig. 3). Some patients also showedelevated levels of IGFBP-2 compared to healthy controls byligand (Fig. 3, Patients 3, 5 and 7) and Western blotting (datanot shown).

Serum IGFBP-3/IGF complexes

Human sera incubated with125I-IGF-II forms multiple 125I-IGF-II/binding protein complexes. When serum was incubatedwith 125I-IGF-II and run on a Sephadex G200 column most ofthe radioactivity eluted in a fraction with a molecular mass of�150 kDa (data not shown). The 150-kDa complex migratedon a non-denaturing PAGE gel in the same position as the majorbinding protein complex in serum from a healthy control(Fig. 4A and B, lane 1). The complex was retarded by IGFBP-3antibodies, but not by control antibodies (Fig. 4B, lane 5, anddata not shown).125I-IGF-II/IGFBP-3 complex formation wasnormal in control sera and in an HIV positive patient who hadhigh IGF levels (Fig. 4B, lanes 1 and 2). In contrast, IGFBP-3complex formation was reduced in a patient with low serumIGF-I levels (Fig. 4B, lane 3). IGFBP-3 complex formation was



Fig. 2 The effect of IGFBP-1 on the abilityof the IGF-I radioimmunoassay to measureIGF-I. Standard curves were run in thepresence and absence of IGFBP-1 at aconcentration that was tenfold higher thanthat contained in sera from patients withAIDS (� BP-1). The ability of acid– ethanolextraction to remove IGFBP-1 from serumwas also evaluated. AIDS patient sera (inset,lane 1), neutralized supernatant from an acid–ethanol extract of AIDS patient sera (lane 2),and the resolubulized pellet from an acidethanol extract (lane 3) were run on a non-denaturing PAGE gel in the presence of125I-IGF-I. Only the starting serum contained anappreciable amount of IGFBP-1.

normal in patients 1,6,8 and 10 (Fig. 4A). These patientshad IGF levels similar to that of healthy control subjects(Fig. 1B). IGFBP-3 complex formation was reduced in patients2–5, 7, 9 and 11. These patients had IGF values that weresignificantly lower than control subjects (compare, Fig. 1b andFig. 4A).

IGFBP-3 proteolysis

Serum from acutely ill (or catabolic) patients often exhibitIGFBP-3 proteolysis. We compared serum IGFBP-3 proteo-lytic activity in a healthy volunteer and 11 AIDS patients.IGFBP-3 protease activity in Patients 4, 7, 8, 10 and 11 wasgreater than that of control serum based on a loss of full-length 125I-IGFBP-3 and the appearance of125I-IGFBP-3proteolytic fragments of 30, 20 and 14 kDa (Fig. 5).125I-IGFBP-3 from these same reactions was found toform ternary complexes on non-denaturing PAGE gels(Fig. 6A). Ternary complex formation was greatest in controlsera and serum from AIDS Patients 1, 2, 6, 8 and 10. Themigration of the IGFBP-3 ternary complex was retarded byaddition of either IGFBP-3 or ALS antibodies, but notantibodies against IGFBP-1 (Fig. 6B).

Restoration of the ternary complex

When125I-IGF-II is added to serum it equilibrates between theternary complex and small IGF binding proteins. By non-denaturing PAGE,125I-IGF-II associated mostly with theternary complex when it was added to control serum (Fig. 7,lane 1). In contrast, in AIDS serum, very little125I-IGF-II wasassociated with the ternary complex. Rather than binding toIGFBP-3,125I-IGF-II bound to IGFBP-1 in serum from patientswith AIDS wasting (Fig. 7, lane 2). Addition of pure ALS toAIDS sera did not change the distribution of125I-IGF-II (Fig. 7,lane 3). Addition of pure IGFBP-3 completely inhibited bindingof 125I-IGF-II to IGFBP-1, but failed to restore the ternarycomplex (Fig. 7, lane 4). Addition of both pure ALS andIGFBP-3 to AIDS sera restored the ternary complex to controllevels (Fig. 7, lane 5).

Serum IGFBP–1

In AIDS patients with wasting, serum IGFBP-1 was present inits most highly phosphorylated form.125I-IGF-I/IGFBP-1complexes co-migrated with pure p4IGFBP-1/125I-IGF- com-plexes I (where p denotes phosphorylated and4 denotesthe migration of the phosphoisoform on non-denaturing



Fig. 3 Serum IGFBP-3 in 11 AIDS patients and three healthy control subjects were compared by ligand blotting. Briefly, 6�l of patient sera waselectrophoresed through a 12.5% SDS-PAGE gel and transferred to nitrocellulose. The resulting blot was probed with125I-IGF-I and II, washed,and used to expose X-ray film. Lanes 1–3 (control), Lanes 4–13 (11 representative AIDS patients). An 18-hour exposure of the resultingautoradiogram is shown. Patient identification is the same as that in Table 1 and Fig. 1. Numbers on the left margin indicate molecular weightmarkers: (44) ovalbumin, (29) carbonic anhydrase, and (18)�-lactoglobulin.

polyacrylamide gels) by non-denaturing PAGE, but not withnon-phosphorylated-IGFBP-1(Fig. 8, lanes 2 and 4). Themigration of IGFBP-1 in AIDS sera was retarded on non-denaturing PAGE gels after addition of IGFBP-1 monoclonalantibodies (Fig. 8, lane 5). Patient sera with the greatest abilityto form 125I-IGF-I/IGFBP-1 complexes had the poorest abilityto form 125I-IGF-II/IGFBP-3 complexes (Patients 3,4,9 and 11;compare Figs 4 and 8). The concentration of IGFBP-1 in AIDSpatients’ sera was almost 4 times higher than in sera of controlpatients who had been fasted for 18 hours (Table 2, Fig. 8 lanes3 and 4).

Patients with high IGFBP-1 and low IGF-I levels exhibitedIGFBP-1 predominantly as a free IGF binding protein (Fig. 9,

lane 3, Patient 11). A patient with high concentrations of IGF-Iin the presence of IGFBP-1 exhibited preformed IGF-I/IGFBP-1 complexes as determined by non-denaturing PAGE andWestern blotting (Fig. 9, lane 1, Patient 6). The addition ofexogenous IGF-I to a serum sample that contained high levelsof IGFBP-1 and low levels of endogenous IGF-I resulted in ashift upwards of p3–4IGFBP-1 on non-denaturing gels (Fig. 9,lane 4, Patient 11). When IGFBP-1 was present in a complexwith IGF-I, additional IGF-I failed to shift the naturallyoccurring IGF-I/IGFBP-1 complex any further (Fig. 9, lane 2,Patient 6). Free IGFBP-1 was identified in all the AIDS patientswho had lost more than 10% of their ideal body mass.Electrophoresis in the presence or absence of IGF-I, followed



Fig. 4 A, IGFBP-3 ternary complexformation was compared in 11 AIDS patientsby incubation of serum with125I-IGF-II followed by non-denaturingPAGE. Lane 1 (control sera), lane 2 (AIDSpatient sera), lane 3 (AIDS patient sera + anti-BP-1), lanes 4–13 (11 representative AIDSpatient sera). Patient identification is the sameas that in Table 1. B, IGFBP-3 complexformation was confirmed by mixing sera froma healthy control subject (lane 1) or AIDSpatients with either a high (lane 2) or low(lane 3) concentration of serum IGF-I with125I-IGF-I. IGFBP-3 was identified by theability of an IGFBP-3 antibody to retard themigration of IGFBP-3 upon non-denaturingPAGE (lane 5). The absence of a bandsuggests a reduced ability to form the IGFBP-3 ternary complex.

by Western blotting, is necessary because it prevents mis-identification of p3 and p4IGFBP-1/IGF complexes as othervariants of IGFBP-1.

Serum GH and IGF-I in patients followed longitudinally

Because 7 of the 11 AIDS patients studied exhibited reducedIGF-I levels, we examined whether these low levels were dueto the cross-sectional nature of our study. We examined agroup of 7 new patients who had been free of opportunisticinfections for up to 25 months. Four of the patients exhibitednormal IGF-I levels (Patients D–G, Table 3). None of thesepatients had lost more than 10% of their IBM. Two of theseven patients lost more than 10% of their IBM either prior toor during the study (Patients A and C). These two patients andone patient who did not lose weight exhibited IGF-I levels lessthan that of control subjects (Fig. 10b, Patients A–C). Twopatients displayed elevated levels of GH at multiple timepoints (Patients A and B, Fig. 10a). The level of phosphory-lated IGFBP-1 was elevated in 2 of the 3 AIDS patients aswell (Fig. 11).

Discussion

Wasting in AIDS is associated with a decreased lean body massand an altered metabolic status (Grunfeld & Feingold, 1992;Baumet al., 1994). Seven of the patients studied in the presentinvestigation lost an average of 15 kg (> 10% IBM). This is thefirst study to demonstrate that AIDS patients exhibit a dramaticdecrease in serum concentrations of IGF-I and IGF-II whenthey have lost more than 10% of their ideal body mass. IGF-Ilevels have previously been found to be normal in paediatricAIDS patients (Laueet al., 1990) and HIV positive adults whohad lost weight (Liebermanet al., 1994). It is likely that ourpatients, given their reduced levels of serum albumin, weremore ill than those studied by Lieberman. Salbeet al.(1991)found that some AIDS patients had lower IGF-I levels thaneither healthy control subjects or AIDS patients who werestable. The present study is the first to show that IGF-IIconcentrations are also decreased in severely wasted patientswith AIDS.

Surprisingly, AIDS patients who displayed the greatestwasting and the lowest serum IGF-I levels also had the highestserum GH concentrations. The combination of elevated GH



Fig. 5 IGFBP-3 proteolysis in serum fromAIDS patients and control subjects. IGFBP-3proteolysis was initiated by mixing125I-IGFBP-3 (25 000 c.p.m.) with 4�l of serumin 0.1 M TRIS buffer at pH 7.4. After an 8-hour incubation at 378C. each reaction wasdivided in half and mixed with sample buffer.To examine proteolysis, the reaction productswere boiled and run on a 12.5% SDS-PAGEgel under non-reducing conditions. A 48-hourexposure of the gel is shown. Molecularweight standards are the same as in Fig. 2.

levels and decreased IGF-I levels suggest a GH resistant state.This conclusion is consistent with the work of Mulliganet al.(1993) who demonstrated that the IGF-I response to exogen-ously administered GH is blunted in AIDS despite thesepatients having normal IGF-I levels. Similar results have beenfound in critically ill patients and patients with GH insensitivitysyndrome (Rosset al., 1991; Laronet al., 1968).

The patients we have studied exhibit many characteristicmanifestations of catabolic patients. Serum cortisol and GHconcentrations were elevated, whereas circulating concentra-tions of insulin, IGF-I and IGF-II were reduced. Although ourpatients were not fasted it is likely that fasting would have

accentuated the differences observed for cortisol, GH, insulinand IGFBP-1 levels between our patients and control subjects.

Since low circulating levels of IGF-I and II could result fromeither reduced hepatic synthesis and/or enhanced clearance ofthe peptides, it was of interest to determine whether AIDSserum contained the major carrier protein for the IGFs. IGFBP-3 was present in serum from AIDS patients as determined byligand blotting, but IGFBP-3 was severely reduced in thosepatients who had lost more than 10% of their ideal body mass.

Our data indicate that IGFBP-3 levels are decreased inseverely wasted AIDS patients and that there are dramaticdifferences in the ability of patient sera to form complexes with



Fig. 6 IGFBP-3 ternary complex formationwas determined on the proteolysis reactionproducts described in Fig. 5. Reactionproducts were run on a 7.5% non-denaturingPAGE gel. A48-hour exposure of the gel is shown.Patients are identified as in Table 1. B,125I-IGFBP-3 ternary complex formation wasconfirmed by the ability of antibodies toindividual components of the complex toretard the migration of the complex. In theabsence of antibody,125I-IGFBP-3 bound toendogenous IGFs and ALS in control serumto form a single band on non-denaturingPAGE (lane 1). Addition of antibody againsteither IGFBP-3 (lane 2) or ALS (lane 3)retarded the migration of the complex. Themigration of the ternary complex wasunchanged by control antibodies againstIGFBP-1 (lane 4).

125I-IGF-II. AIDS patients who displayed low serum concen-trations of IGF-I and -II showed the poorest ability to formIGFBP-3 ternary complexes on non-denaturing PAGE gels.These patients also lacked preformed ternary complexes asdetermined by non-denaturing PAGE and Western blotting(data not shown). IGFBP-3 may be proteolysed to a form that isno longer capable of forming a ternary complex or, alter-natively, AIDS sera may be deficient in ALS.

We found that AIDS patients displayed IGFBP-3 proteolyticactivity in their serum, but that the level of proteolytic activitywas not necessarily associated with low serum IGF levels orwastingper se. In contrast, the ability of125I-IGFBP-3 to formternary complexes with IGFs and ALS in individual serumsamples did correlate with serum IGF concentrations. Ternarycomplex formation was greatest in patients with high IGF-Ilevels (patients 1, 2, 6, 8 and 10). This agrees with data weobtained using125I-IGF-II to assess ternary complex formation.This also suggests that the ternary complex is stable in patientsera for more than 8 hours at 378C.

The reduced ability of serum from AIDS patients to form theternary complex may lead to reduced levels of circulating IGF-Iand IGF-II over the course of the patient’s illness. A similarreduction in IGF-I and IGF-II has been reported to occur inchildren with shigellosis (Pucilowskaet al., 1993). It is possiblethat malnutrition or poor liver function may account for some ofthe changes in the present study. However, all patients had acalorie intake that met or exceeded their estimated restingenergy expenditure. Patients also produced high levels of serumIGFBP-1. This indicates that low IGF-I levels are not a result ofa generalized hepatic failure.

IGFs have been hypothesized to equilibrate between theternary complex and small IGFBPs in serum (Smithet al.,1995). Since normal control serum contains a molar excess ofIGFBP-3 and IGFBP-3 has a much higher affinity for IGF-Ithan does IGFBP-1, almost all of the IGFs associate withIGFBP-3. In marked contrast, IGFs preferentially associatedwith IGFBP-1 in serum from patients with wasting. This waslargely due to low levels of IGFBP-3, since addition of IGFBP-3 to AIDS serum prevented125I-IGF-II from binding to IGFBP–1. Although IGFBP-3 bound125I-IGF-II, it failed to form aternary complex. Additional ALS was necessary for complexformation. The distribution of IGFs in patients with AIDSwasting has a striking resemblance to patients who havereceived long-term treatment with the somatostatin analogueoctreotide (Barrecaet al., 1995).

Clinical studies have found IGF-I to have its greatestanabolic effect when it is present in supraphysiologicalconcentrations as part of the IGFBP-3 ternary complex(Kupfer et al., 1993). Low levels of functional IGFBP-3 mayshorten the half-life of IGF-I. Low levels of IGFBP-3 have alsobeen hypothesized to account for the only transient ability ofIGF-I to improve nitrogen balance in AIDS (Liebermanet al.,1994).

AIDS patients with low serum IGF concentrations also hadelevated levels of IGFBP–1. Moreover, IGFBP-1 was present inits most highly phosphorylated form. We have previouslyshown that IGFBP-1 is present as a phosphoprotein in two othercatabolic states, trauma and uncontrolled diabetes mellitus(Wojnaret al., 1995; Frostet al., 1994). High levels of IGFBP-1



Fig. 7 The ability of pure IGFBP-3 and ALS to affect thedistribution of125I-IGF-II between the ternary complex and IGFBP-1was examined.125I-IGF-II was mixed with either control serum (lane1) or serum from an AIDS patient with more than 10% wasting(lanes 2–5). The affect of pure ALS (lane 3), IGFBP-3 (lane 4) orboth proteins (lane 5) was examined by non-denaturing PAGE.Ternary complex was identified as in Figs 4B and 6B. IGFBP-1 wasidentified as in Fig. 8.

may be a result of elevated serum cortisol and/or theinsulinopaenia that often characterize hypermetabolic condi-tions. In our study, IGFBP-1 was present mostly as a free formin patients with low levels of IGF-I (< 100�g/l). In contrast,AIDS patients with high levels of IGF-I (> 150�g/l) displayednative IGF-I/IGFBP-1 complexes by non-denaturing PAGE andWestern blotting. Although such complexes have beenhypothesized to exist (Leeet al., 1993), this is the first timethat they have been adequately demonstrated.

IGF-I may be rapidly bound by IGFBP-1 within the liver,since both proteins are secreted by hepatocytes (Leeet al.,1993). This is in contrast with IGFBP-3 which is synthesizedprimarily by sinusoidal endothelial cells (Chinet al., 1994).The anatomical compartmentalization of these cell types mayfavour the formation of binary complexes with IGFBP-1 ratherthan a ternary complex with IGFBP–3.

IGFs bound to IGFBP-1 may be rapidly cleared (Gargoskyetal., 1993) or the complex may be transported across thecapillary endothelium to the extracellular fluid compartment(Bar et al., 1990). An important consideration is what form of

IGFBP-1 is present in extracellular fluid and whether it is freeor bound to IGF-I. The assessability of IGFs that are bound tophosphorylated IGFBP-1 in tissues remains to be determined.Unlike non-phosphorylated IGFBP-1 (Jyunget al., 1994),phosphorylated IGFBP-1 has predominantly inhibitory effectsboth in vitro and in vivo (Lee et al., 1993; Joneset al., 1991;Frostet al., 1993). We therefore speculate that the increase inphosphorylated IGFBP-1 in AIDS patients would result indiminished IGF availability and action.

The cross-sectional nature of the first part of our study led usto determine whether the GH/IGF-I axis is abnormal forprolonged periods of time in AIDS patients. Longitudinal datain AIDS patients might exclude the possibility that we wereobserving only transient changes in the GH/IGF-I axis thatoccur with episodes of infection or malnutrition (Salbeet al.,1991). In seven patients who were studied for durations up to 25months we found that four had normal IGF-I levels (210�26�g/l). These patients maintained their body weight throughout thestudy. Two of the remaining three patients had lost 10% ormore of IBM prior to or during the study. These two patients



Fig. 8 IGFBP-1 levels in 11 AIDS patient sera were compared to a control patient fasted for 18 hours (control). Native125I-IGF-I/IGFBP-1complexes were formed by addition of125I-IGF-I to 6 �l of patient sera or pure IGFBP-1 standards. The complexes were electrophoresed througha non-denaturing polyacrylamide gel. IGFBP-1 was identified by addition of an IGFBP-1 monoclonal antibody (clone 6303) which retards themigration of the IGF-I/IGFBP-1 complex. Lane 1 (non-phosphorylated IGFBP-1), lane 2 (phosphorylated IGFBP-1), lane 3 (control sera), lane 4(AIDS sera), lane 5 (AIDS patient sera + Anti-BP1), Lanes 6–16 (11 representative AIDS patient sera). The presence and intensity of the band(s)corresponds to the relative amount of free IGFBP-1 in each sample. A 15-hour exposure of the dried gel is shown. Patient identification is thesame as that in Table 1.

showed IGF-I levels that were�50% of healthy volunteers.IGF-I levels remained low in these patients for up to 25 months.GH and IGFBP-1 were also abnormally high at many of thetime points examined. A subset of AIDS patients thereforeappear to display persistent changes in their GH/IGF axis. Sincepatients who were followed longitudinally did not have acuteopportunistic infections, our data suggest that mechanisms existthat prevent restoration of IGF levels to normal. These may

include the abnormal expression and post-translational mod-ification of IGFBPs. Although AIDS associated wasting isundoubtedly a multifactorial process, the present studyidentified four changes that occur in the GH/IGF axis thatmay contribute to a reduced anabolic potential in AIDS patientsthat have lost more than 10% of their IBM. The changesinclude: (a) reduced levels of circulating IGF-I, IGF-II, andIGFBP-3; (b) an inability of high concentrations of GH tomaintain normal levels of IGF-I; (c) a reduced ability ofIGFBP-3 to form ternary complexes with IGF-I and IGF-II; and(d) a marked increase in a free and phosphorylated form ofserum IGFBP-1. Furthermore, these changes are not merelytransient but are sustained.

In the future it will be important to determine whatprecipitates the observed changes in the serum IGF systemand at what stage of HIV infection these changes occur. It ishoped that therapies can be devised for reversing or slowingAIDS-associated wasting, perhaps by preventing changes in theGH/IGF-I axis.

Acknowledgements

We thank Ken Marsh for helping in the preparation ofiodinated peptides and Jie Fan for performing the GH RIA.We thank Dr Robert Baxter for kindly providing antisera tothe acid labile subunit of the IGFBP-3 ternary complex.



Fig. 9 The form of IGFBP-1 in AIDSpatients 6 and 11 was compared by non-denaturing PAGE followed by Westernblotting. Patient number and the presence orabsence of exogenous IGF-I (250 ng) islabelled above each lane. The relativemigration of free and bound IGFBP-1 isnoted. Note that 6-times more serum wasloaded on the gel for patient 6 than forpatient 11 in order to achieve equivalentsignals by Western blotting.

Table 3 Serum IGF-1 in AIDS patients followed longitudinally

IGF-l % IBMPatient (�g/l) Points Months lost

A 97� 13 6 18 10B 81� 22 6 19 0C 101� 32 4 25 11D 248� 71 6 18 0E 193� 43 9 20 0F 202� 75 9 20 0G 383� 162 6 15 4

Values are means�SEM.‘Points’ represents the number of samples collected per patient over thenumber of ‘months’ specified.% IBM Lost is the amount of weight lost during the study period, or theamount of weight lost prior to the collection of the first sample.

This work was supported by NIH grant GM38032 (CHL),DK49316-01 (MCG) and an NIH-sponsored InstitutionalResearch Service Award T32DK07521 to the Diabetes andMetabolic Diseases Research Program-Stony Brook (RAF).

References

Baum, M., Cassetti, L., Bonvehi, P., Shor-Posner, G., Lu Y, andSauberlich, H. (1994) Inadequate dietary intake and altered nutritionstatus in early HIV-1 infection.Nutrition, 10, 16–20.

Baxter, R.C., Hizuka, N., Takano, K., Halman, S.R. & Asakawa, K.(1993) Responses of insulin-like growth factor binding protein-1(IGFBP-1) and the IGFBP-3 complex to administration of insulin-like growth factor-I.Acta Endocrinologica, 128,101–108.

Bereket, A., Lang, C.H., Blethen, S.L., Fan, J., Frost, R.A. &Wilson, T.A. (1995) Insulin-like growth factor binding protein-3proteolysis in children with insulin-dependent diabetes mellitus: apossible role for insulin in the regulation of IGFBP-3 proteaseactivity. Journal of Clinical Endocrinology and Metabolism,80,2282–2288.

Chin, E., Zhou, J., Dai, J., Baxter, R.C. & Bondy, C.A. (1994) Cellularlocalization and regulation of gene expression for components of theinsulin-like growth factor ternary complex.Journal of ClinicalEndocrinology and Metabolism, 134,2498–2504.

Clemmons, D.R., Camcho-Hubner, C., Jones, J.I., McCusker, R.H. &Busby, W.H. (1991) Insulin like growth factor binding proteins:mechanism of action at the cellular level. InModern Concepts ofInsulin-like Growth Factors (ed. E.M.Spencer), pp. 475–486,Elsevier, New York.

Clemmons, D.R., Smith-Banks, A. & Underwood, L.E. (1992)Reversal of diet- induced catabolism by infusion of recombinantinsulin-like growth factor-I in humans.Journal of ClinicalEndocrinology and Metabolism, 75, 234–238.

Clemmons, D.R., Snyder, D.K., Williams, R. & Underwood, L.E.(1987) Growth hormone administration conserves lean bodymass during dietary restriction in obese subjects.Journal ofClinical Endocrinology and Metabolism, 64, 878–883.

Cohick, W.S. & Clemmons, D.R. (1993) The insulin like growthfactors.Annual Review of Physiology, 55, 131–153.

Conover, C.A., Kiefer, M.C. & Zapf, J. (1993) Posttranslationalregulation of insulin-like growth factor binding protein-4 in normaland transformed human fibroblasts. Insulin-like growth factor depend-ence and biological studies.Journal of Clinical Investigation, 91,1129–1137.

Dahn, M.S., Lange, M.P. & Jacobs, L.A. (1988) Insulin like growthfactor-I production is inhibited in human sepsis.Archives of Surgery,123,1409–1414.

Frost, R.A., Berekett, A., Wojnar, M.M., Wilson, T.A.,Lang, C.H. & Gelato, M.C. (1994) Phosphorylation of humaninsulin-like growth factor binding protein-1 in patients with



Fig. 10 IGF-I and GH levels were measured in 8 patients over a 25-month period (Table 3).a, IGF-I and b, GH concentrations are presented for 3of the patients who displayed low IGF-I levels (Patients A, B and C).

Fig. 11 The level of phosphorylated IGFBP-1 in patients A,B and C was followed longitudinally as determined by non-denaturing PAGE with125I-IGF-I. Patients A, B and C and the assay time points correspond to those described in Fig. 10.

insulin-dependent diabetes mellitus and severe trauma.Journal ofClinical Endocrinolology and Metabolism, 78, 1533–1535.

Frost, R.A., Mazella, J. & Tseng, L. (1993) Insulin-like growth factorbinding protein-1 inhibits the mitogenic effect of insulin-like growthfactors and progestins in human endometrial stromal cells.Biology ofReproduction, 49, 104–111.

Frost, R.A. & Tseng, L. (1991) Insulin-like growth factor bindingprotein-1 is phosphorylated by cultured human endometrial stromalcells and multiple protein kinasesin vitro. Journal of BiologicalChemistry, 266,18082–18088.

Fryburg, D.A., Barrett, E.T. & Gelfand, R.A. (1991) Growth hormoneacutely stimulates forearm muscle protein synthesis in normalhumans.American Journal of Physics,260,E499-E504.

Gargosky, S.E., Hasegawa, T., Tapanainen, P., MacGillivray, M.,Hasegawa, Y. & Rosenfeld, R.G. (1993) Urinary insulin-like growthfactors (IGF) and IGF-binding proteins in normal subjects, growthhormone deficiency, and renal disease.Journal of ClinicalEndocrinology and Metabolism, 76, 1631–1637.

Gelato, M.C., Vasualotti, J., Spatola, E., Rutherford, C., Marsh, K. &Carlson, H.E. (1992) Differential tissue regulation of the insulin-likegrowth factors in rats bearing the MstT/W15 pituitary tumor.Neuroendocrinology, 56, 765–774.

Grinspoon, S.K. & Bilezikian, J.P. (1992) HIV disease and the endo-crine system.New England Journal of Medicine, 327,1360–1365.

Grunfeld, C. & Feingold, K.R. (1992) Metabolic disturbances andwasting in the acquired immunodeficiency syndrome.New EnglandJournal of Medicine, 327,329–337.

Jones, J.I., D’Ercole, A.S., Camacho-Hubner, C. & Clemmons, D.R.(1991) Phosphorylation of insulin-like growth factor binding protein-1 in cell culture andin vivo: effects on affinity for IGF-I.Proceedingsof the National Academy of Science of the USA, 88, 7481–7485.

Jyung, R.W., Mustoe, T.A., Busby, W.H. & Clemmmons, D.R. (1994)Increased wound breaking strength induced by insulin-like growthfactor I in combination with insulin-like growth factor bindingprotein–1.Surgery, 115,233–239.

Kupfer, S.R., Underwood, L.E., Baxter, R.C. & Clemmons, D.R. (1993)Enhancement of the anabolic effects of GH and insulin-like growthfactor-I by use of both agents simultaneously.Journal of ClinicalInvestigation, 93, 391–396.

Langfod, K.S. & Miell, J.P. (1993) The insulin-like growth factor-1/binding protein axis: physiology, pathophysiology and therapeuticmanipulation.European Journal of Clinical Investigation, 23, 503–516.

Laron, Z., Pertzelan, A. & Karp, M. (1968) Pituitary dwarfism with highserum levels of growth hormone.Israel Journal of Medical Science,4, 883–894.

Lave, L., Pizzo, P.A., Butler, K. & Cutler, G.B. (1990) Growth andneuroendocrine dysfunction in children with acquiredimmunodeficiency.Journal of Pediatrics, 117,541–545.

Lee, P.D.K., Conover, C.A. & Powell, D.R. (1993) Regulation andfunction of insulin-like growth factor-binding protein–1.Proceedingsof the Society for Experimental Biology and Medicine, 204,4–29.

Lieberman, S.A., Butterfield, G.E., Harison, D. & Hoffman, A.R.(1994) Anabolic effects of recombinant insulin-like growth factor-Iin cachectic patients with the acquired immunodeficiencysyndrome. Journal of Clinical Endocrinology and Metabolism,78, 404–410.

Mulligan, K., Grunfeld, C., Hellerstein M.K. Neese, R.A. &Schambelanm, M. (1993) Metabolic effects of recombinant humangrowth hormone in patients with wasting associated with humanimmunodeficiency virus infection.Journal of Clinical Endocrinol-ogy and Metabolism, 77, 956–962.

Olchovsky, D., Gelato, M.C., Bruno, J.F., Song, J. & Berelowitz, M.(1991) Pituitary IGF-I content and gene expression in thestreptozotocin diabetic rat: evidence for tissue specific regulation.Endocrinology, 128,923–928.

Press, M. (1988) Growth hormone and metabolism.Diabetes/Metabolism Reviews, 4, 391- 414.

Pucilowska, J.B., Davenport, M.L., KIabir, I., Clemmons, D.R.,Thissen, J.P., Butler, T. & Underwood, L.E. (1993) The effectof dietary protein supplementation on insulin-like growthfactors (IGFs) and IGF-binding proteins in children withshigellosis.Journal of Clinical Endocrinology and Metabolism, 77,1516–1521.

Roeder, R.A., Hossner, K.L., Sasser, R.G. & Gunn, J.M. (1986)Regulation of protein turnover by recombinant insulin-like growthfactor-I in L6 myotube cultures.Hormone and Metabolic Research,20, 698–700.

Ross, R., Miell J, and Freeman, E. (1991) Critically ill patients havehigh basal growth hormone levels with attenuated oscillatory activityassociated with low levels of insulin-like growth factor-I.ClinicalEndocrinology, 35, 47–54.

Salbe, A.D., Kotler, D.P., Wang, J., Pierson, R.N. & Campbell, R.G.(1991)Clinical Research, 39, 385 A.

Salomaon, F., Cuneo, R. & Sonksen, P.H. (1991) Growth hormone andprotein metabolism.Hormone Research, 36, 41–43.

Shimasaki, S. & Ling, N. (1992) Identification and molecularcloning of insulin-like growth factor binding proteins (IGFBP-1,-2, -3, -4, -5, and -6).Progress in Growth Factor Research, 3,243–266.

Shimizu, M., Webster, C., Morgan, D.O., Blau, H.M. & Roth, R.A.(1986) Insulin and insulin-like growth factor-I receptors andresponses in cultured human muscle cells.American Journal ofPhysiology,251, E611-E615.

Smith, W.J., Underwood, L.E., Clemmons, D.R. (1995) Effectof caloric or protein restriction on insulin-like growth factor-I(IGF-I) and IGF-binding proteins in children and adults.Journal ofClinical Endocrinology and Metabolism, 80, 443–449.

Stein, J.P., Nutrinsky, C., Condoluci, D., Schluter, M.D. & Leskiw, M.J.(1990) Protein and energy substrate metabolism in AIDS patients.Metabolism, 39, 876- 8811.

Wojnar, M.M., Fan, J., Frost, R.A., Gelato, M.C. & Lang, C.H. (1995)Alterations in the insulin-like growth factor system in traumapatients.American Journal of Physiology,268, R970-R977.



Wasting in the acquired immune deficiency syndrome is associated with multiple defects in the serum...

Documents

Transcript of Wasting in the acquired immune deficiency syndrome is associated with multiple defects in the serum...