Expe

rim

enta

lPhy

siol

ogy

Exp Physiol 00.00 (2012) pp 1–11 1

Research PaperResearch Paper

Time course of hepatic gluconeogenesis during hindlimbsuspension unloading

Ilya R. Bederman1, Visvanathan Chandramouli2, Yana Sandlers3, Leigh Henderson1 and Marco E. Cabrera1,2

Departments of 1Pediatrics, 2Endocrinology, 3Nutrition and 4Biomedical Engineering, Case Western Reserve University, Cleveland, OH 44106, USA

The goal of this work was to determine the time-dependent changes in fractional hepaticgluconeogenesis (GNG) during conditions of hindlimb suspension unloading (HSU), a ‘ground-based’ method for inducing muscular atrophy to simulate space flight. We hypothesized thatGNG would increase in HSU conditions as a result of metabolic shifts in the liver and skeletalmuscle. A significant and progressive atrophy was observed in the soleus (30, 47 and 55%) andgastrocnemius muscles (0, 15 and 17%) after 3, 7 and 14 days of HSU, respectively. Fractionalhepatic GNG was determined following the incorporation of deuterium from deuteratedwater (2H2O) into C–H bonds of newly synthesized glucose after an 8 h fast. Enrichmentof plasma glucose with 2H was measured using the classic method of Landau et al. (the‘hexamethylenetetramine (HMT) method’), based on specific 2H labelling of glucose carbons,and the novel method of Chacko et al. (‘average method’), based on the assumption of equal2H enrichment on all glucose carbons (except C2). After 3 days of HSU, fractional GNG wassignificantly elevated in the HSU group, as determined by either method (∼13%, P < 0.05). After7 and 14 days of HSU, gluconeogenesis was not significantly different. Both analytical methodsyielded similar time-dependent trends in gluconeogenic rates, but GNG values determined usingthe average method were consistently lower (∼30%) than those found by the HMT method. Tocompare and validate the average method against the HMT method further, we starved animalsfor 13 h to allow for hepatic GNG to contribute 100% to endogenous glucose production. TheHMT method yielded 100% GNG, while the average method yielded GNG of ∼70%. As bothmethods used the same values of precursor enrichment, we postulated that the underestimationof gluconeogenic rate was as a result of differences in the measurements of product enrichment(2H labelling of plasma glucose). This could be explained by the following factors: (i) loss ofdeuterium via exchange between acetate and glucose; (ii) interference caused by fragment m/z169, representing multiple isobaric species; and (iii) interference from other sugars at m/z 169.In conclusion, HSU caused a time-dependent increase in hepatic gluconeogenesis, irrespectiveof the analytical methods used.

(Received 1 May 2012; accepted after revision 15 June 2012; first published online 15 June 2012)Corresponding author I. R. Bederman: Case Western Reserve University, 10900 Euclid Avenue, BRB/830, Cleveland,OH 44106, USA. Email: ilya@case.eduM. E. Cabrero is sadly deceased.

Space exploration is associated with a lack of thegravitational field to which humans have adapted onEarth. Thus, during weightlessness, physiological changessuch as fluid shifts, bone desorption, muscular atrophyand endocrine changes occur in the human body.To study the effects of zero gravity on physiologyand metabolism, a rodent model of ‘ground-based’

weightlessness, i.e. hindlimb suspension unloading (HSU)has been established (Musacchia et al. 1983). DuringHSU, rodents are suspended at an angle that ‘unloads’their hindlimbs, resulting in significant muscular atrophy(Morey-Holton et al. 2005) and alterations in substratemetabolism, e.g. increased glycolysis, decreased fatoxidation and accumulation of triglycerides and glycogen

C© 2012 The Authors. Experimental Physiology C© 2012 The Physiological Society DOI: 10.1113/expphysiol.2012.067074

2 I. R. Bederman and others Exp Physiol 00.00 (2012) pp 1–11

(Thomason & Booth, 1990; Stein & Wade, 2005). Inresponse to HSU, metabolic changes occur in the liver.Stein & Wade (2005)) reported significant increases inmRNA and activity of hepatic gluconeogenic enzymes,suggesting that hepatic gluconeogenesis (GNG) maybe increased. However, no direct measurement ofgluconeogenic rate during conditions of HSU has beenpublished to date. Thus, the goal of the work presentedhere was to quantify fractional hepatic gluconeogenesisafter 3, 7 and 14 days of HSU.

We determined fractional hepatic gluconeogenesisusing incorporation of deuterated water (2H2O) into C–Hbonds of newly made glucose. Briefly, after either an 8 or a13 h fast, rats were given a bolus of 2H2O to enrich the bodywater with deuterium. Gluconeogenesis was determinedfrom analyses of 2H labelling of plasma glucose, whichwere done using two methods for comparison andvalidation. First, we determined gluconeogenesis usingclassic methodology developed and validated by Landauand colleagues (Chandramouli et al. 1997), commonlyreferred to as the ‘hexamethylenetetramine (HMT)method’. Second, we analysed plasma samples usingnovel methodology, recently developed by Chacko et al.(2008a)). Briefly, the HMT method is based on the notionthat 2H from body water incorporated into carbon 5(C5) of newly synthesized glucose was shown by Landauand colleagues to be specific for hepatic gluconeogenesis(Chandramouli et al. 1997). Incorporation of 2H intocarbon 2 (C2) of glucose represents both gluconeogenesisand glycogenolysis and equals that of body water atsteady state (Landau et al. 1995b, 1996; Chandramouliet al. 1997). Consequently, the ratio of (2H enrichmentof C5)/(2H enrichment of body water or C2) representsfractional gluconeogenesis. The HMT method offersmajor advantages over other techniques (Previs &Brunengraber, 1998). First, administered deuterated wateris quickly distributed and fully equilibrated with thebody water, thus exposing all glucose-synthesizing cellsto the same isotopic enrichment. This takes care of anypossible issues of heterogeneity of precursor enrichment(e.g. liver zonation) that occur when 13C substratesare utilized (Landau et al. 1995a; Previs et al. 1995,1998). Second, deuterated water is a relatively inexpensive,safe and versatile isotope. The versatility of deuteriumincorporation has permitted development of methods ofmeasuring de novo synthesis of lipids, sterols, proteinsand carbohydrates in humans and animals (Previs &Brunengraber, 1998; Katanik et al. 2003; McCabe & Previs,2004; Dufner et al. 2005; Bederman et al. 2006). Third,sixfold amplification of C5 2H labelling permits a lowdose of 2H2O to be administered. One of the disadvantagesof the method is the multistep purification and isolationof C5 of glucose (Schumann et al. 2001). Nevertheless,Landau’s method has been widely applied in many humanand animal studies (for review, see Previs & Brunengraber,1998; McCabe & Previs, 2004).

The analytical complexity of the HMT method ledto the development of a simpler approach by Chackoet al. (2008a). Similar to the HMT method, Chacko’smethod is based on the incorporation of deuteriumfrom body water into newly synthesized hepatic glucoseduring gluconeogenesis. The main assumption of themethod is that all the carbons of glucose apart fromC2 have equal labelling distribution of the incorporated2H atoms. Thus, isolation of specific carbons, i.e. C5 ofglucose, is not necessary. Fractional gluconeogenesis isthen computed by dividing total 2H labelling of plasmaglucose by 6. As the method is based on the averageglucose labelling, we will refer to it as the ‘average method’throughout this manuscript. The 2H enrichment of plasmaglucose is determined simply by extracting plasma glucoseusing cold acetone and direct peracetylation. Using gaschromatography–mass spectrometry (GC–MS), chemicalionization and isotopic shifts of known deuteratedstandards, Chacko et al. (2008a) elucidated that theglucose fragment m/z 169 represented the intact backboneof the molecule. The ratio m/z 170/169 was then used todetermine product enrichment and calculate fractionalgluconeogenesis using the precursor–product relationshipas in the HMT method. We were attracted to utilizeChacko’s method for the following reasons: (i) it usesdeuterium incorporation from body water enriched with2H2O and thus would not require additional animals orsamples; (ii) analytical simplicity; and (iii) demonstratedsimilarity to the HMT method. We compare data obtainedusing both the HMT and the average analytical methodsand discuss the differences herein.

Methods

Chemicals and supplies

Unless specified, all chemicals and reagents werepurchased from Sigma-Aldrich. Deuterated water waspurchased from Isotec (Miamisburg, OH, USA). Tinctureof Benzoin

R©was purchased from Fisher Scientific,

(Pittsburgh, PA, USA).

Biological experiments

All animal care and use was approved by the InstitutionalAnimal Care and Use Committee of Case Western ReserveUniversity. The HSU protocols were drafted using NASA-based recommendations and approved by the university’sveterinarian. Young Sprague–Dawley male rats, weighing∼200 g, were purchased from Taconic Farms (Hudson,NY, USA). After a brief acclimation to the animal facility,animals were split into two groups for studies I and II.Within each study, rats were further split to control (C)and hindlimb-suspended (HSU) groups, with n = 4 inevery group.

C© 2012 The Authors. Experimental Physiology C© 2012 The Physiological Society

Exp Physiol 00.00 (2012) pp 1–11 Gluconeogenesis during hindlimb suspension 3

Hindlimb suspension unloading. Individual HSU cageswere custom made from clear acrylic panels (0.22 mm)based on published designs (Wronski & Morey-Holton,1987; Morey-Holton et al. 2005). Cages were of∼1032 cm2, which exceeds the minimal recommendedfloor space requirement of 452 cm2 per animal. The cagefloor was made from ‘eggcrate’ grid panel as suggestedby Park & Schultz (1993) to allow for animal wasteto fall through. The suspension apparatus consisted ofa swivel pulley, which glided on the smooth stainless-steel rod affixed at the top of the cage, thus allowing for360 deg animal movement. A water bottle was affixed tothe cage wall by a spring. Food pellets were in a smalldish affixed to the cage floor. To affix the animal to thesuspension apparatus, the rat was first sedated by briefinhalant anesthesia (isoflurane) and the entire tail wascleaned with Tincture of Benzoin

R©. After the tincture

dried and became sticky, traction tape was glued to two-thirds of the tail with non-irritant adhesive. The tape wasthen attached to a large paper clip by steel staples. Theglue was allowed to dry for 15 min while the rat recoveredfrom anesthesia and was attached to the swivel pulley. Theheight of the suspension apparatus was adjusted as neededto maintain the recommended 30 deg angle with the floor.At this angle, the animal was able to rest on the front half ofits body comfortably and to walk around on its front paws.Animal health status was carefully monitored on a dailybasis. Animal body weight, food and water intakes wererecorded daily. Animals were suspended for 24 h day−1 for3, 7 or 14 days.

Study I (short-term fasting). Animals from the controlgroup remained in normal cages, while the HSUgroup underwent the HSU procedure and remained insuspension for 3, 7 or 14 days. On days 4, 8, 15, food wasremoved at 07.00 h in both the control and the HSU group.After 6 h of fasting (13.00 h) on each of the respectivedays, rats were injected with 2H2O (see ‘Administration ofdeuterated water’ below).

Study II (long-term fasting). Animals from the controlgroup remained in normal, cages while the HSUgroup underwent the HSU procedure and remained insuspension for 14 days. On day 14, food was removed at22.00 h in both the control and the HSU group. After 11 hof fasting (09.00 h), rats were injected with 2H2O (see nextsubsection).

Administration of deuterated water. Animals receivedan intraperitoneal injection of 22 μl 2H2O-saline (g bodyweight)−1 (99.7% atom excess 2H2O + 0.09 g l−1 NaCl)under brief inhalant general anaesthesia. Note that thisdose of 2H2O should yield∼3.0% 2H-enriched body water.For studies I and II, 2 h postinjection the rats were given

terminal anaesthesia via an overdose of inhalant anestheticisoflurane and blood was removed via direct cardiacpuncture, at 15.00 and 11.00 h, respectively. Plasma wasremoved by centrifugation at 2770 g and stored at −20◦C.Soleus, extensor digitorum longus (EDL), and mixedgastrocnemius (GC) muscles were quickly excised, blotteddry, weighed, snap-frozen in liquid nitrogen, and storedat −80◦C. The total fasting times were 8 (short term) and13 h (long term).

Analytical procedures

The 2H labelling of plasma water was determined by themodified acetone exchange procedure originally describedby Yang et al. (1998). Briefly, 20 μl of sample or standardwas mixed with 3 μl of 10 N NaOH and 3 μl of a 5%(v/v) solution of acetone in acetonitrile, centrifuged andreacted at room temperature overnight. Acetone was thenextracted by the addition of 600 μl of chloroform, followedby addition of anhydrous sodium sulfate. Samples werevigorously mixed, and a small aliquot of chloroform wastransferred to a GC–MS vial.

The HMT method

To determine fractional gluconeogenesis, we measured2H labelling on C2 and C5 of glucose (Chandramouliet al. 1997; Schumann et al. 2001). Briefly, plasma sampleswere first deproteinized by addition of 0.3 N ZnSO4 andBa(OH)2. The supernatant was then deionized by using amixture of ion exchange resin beds consisting of AG1-X8 formate and AG 50W-X8 in the H+ form. Theglucose in the effluent was purified by HPLC as published(Schumann et al. 2001). Purified glucose was diluted withunlabelled glucose to reach 0.5% 2H labelling and used todetermine C2 and C5 2H labelling of glucose. For C2 2Hlabelling, glucose was enzymatically converted to a mixtureof ribulose-5-phosphate and arabitol-5-phosphate asdescribed previously (Landau et al. 1995b). The mixturewas oxidized with periodic acid, yielding formaldehydethat contained C2 2H labelling. The latter was thenreacted with ammonia to yield hexamethylenetetramine(HMT). For C5 labelling, glucose was converted to xyloseas described previously (Schumann et al. 2001). 1,2-O-Isopropylidene-D-xylofuranose, an intermediate of xyloseconversion from glucose, was directly hydrolysed with0.1 N H2SO4 omitting the ethyl acetate extraction. Theresulting xylose was oxidized with periodate, yieldingformaldehyde that contained C5 2H labelling. Again, asin the case of C2, the formaldehyde containing C5 2Hlabelling was reacted with ammonia to yield HMT. Eachsample was injected in triplicate in GC–MS.

Gas chromatography–mass spectrometry. In allanalyses, we used an Agilent 5973N-MSD equipped

C© 2012 The Authors. Experimental Physiology C© 2012 The Physiological Society

4 I. R. Bederman and others Exp Physiol 00.00 (2012) pp 1–11

with an Agilent 6890 GC system (Agilent, Santa Clara,CA, USA) that contained a DB-17MS capillary column(30 m × 0.25 mm × 0.25 μm). The 2H labelling of HMTwas determined as published elsewhere (Schumann et al.2001). The injector port was held at 220◦C. The initialtemperature programme was set at 100◦C and held for2 min, increased by 20◦C min−−1 to 220◦C, with a heliumflow of 1 ml min−1. Selective ion monitoring of m/z 140and 141 was used. Standards of HMT prepared from 0–2.0% 2H enrichment of (1-2H)sorbitol were used for datacorrection.

Calculations. The fractional gluconeogenesis wascalculated by dividing C5 2H enrichment by either C2or body water 2H enrichment to check the validity of bodywater equilibration with C2 of glucose.

The average method

Plasma samples were first deproteinized by adding 100μl of cold acetone. The supernatant was evaporatedto dryness, and glucose was acetylated using 50 μl ofacetic anhydride/pyridine (2:1) mixture overnight at roomtemperature. The sample was evaporated to dryness andreconstituted in ethyl acetate, and a 1 μl aliquot wasinjected into GC–MS in triplicate.

Gas chromatography–mass spectrometry. The 2Hlabelling of the glucose penta-acetate derivative wasdetermined using an Agilent 5973N-MSD equipped withan Agilent 6890 GC system. A VF-5MS capillary column(30 m × 0.25 mm × 0.32 μm) was used for this assay. Theinjector port was held at 250◦C. The initial temperatureprogramme was set at 70◦C and held for 1 min, increasedby 30◦C min−1 to 275◦C, with a helium flow of 1 ml min−1.The mass spectrometer was operated in the chemicalionization mode using methane as the reagent gas.Selective ion monitoring was done using m/z 169 and170. Dwell time was set at 10 ms.

Calculations. The fractional gluconeogenesis wasdetermined using the precursor–product relationship.Molar precursor enrichment (2H labelling of bodywater; MPE) was determined as outlined in the‘Analytical procedures’ section, via acetone exchange.Product enrichment was determined assuming that allexchangeable hydrogens of glucose except C2 (near equalto body water 2H labelling) have equal 2H enrichment.Thus, per atom 2H enrichment of glucose (productenrichment), i.e. average (M + 1)d, was calculated asfollows:

Average (M + 1)d = (M + 1)d(m/z 169)

/6 (1)

Table 1. Body weight changes during the studies

Control group HSU group

DayInitial

weight (g)Final

weight (g)Initial

weight (g)Final

weight (g)

3 274 ± 9.0 283.3 ± 9.4∗ 290.3 ± 8.8 283.0 ± 10.07 300.7 ± 8.4 330.0 ± 9.4∗ 303.3 ± 3.9 312.7 ± 3.114 291.7 ± 3.1 348.3 ± 2.8∗ 284.0 ± 8.4 298.7 ± 13.7

Data obtained on day 14 are from both studies combined.Data are means ± SEM (n = 4). Initial weights were measuredon day 0, before hindlimb suspension unloading (HSU) began.Final weights were measured on the respective days beforeexsanguination. Significance was determined using Student’sunpaired t test (∗P < 0.05 compared with initial weight).

where (M + 1)d(m/z 169) is the matrix-corrected 2Henrichment of M + 1 from the 170/169 ratio. Fractionalgluconeogenesis was then calculated as follows:

Fractional GNG = Average (M + 1)d/

MPEbody water

(2)

where MPEbody water was determined as described inBederman et al. (2006).

Results

Sprague–Dawley rats were subjected to hindlimbsuspension unloading for 3, 7 and 14 days to simulatemicrogravity. As HSU is designed to restrict the movementof animals by attaching the hindlimbs at a shallow angleabove the floor, this unnatural position imposes sometransient stress. As a result, body weight was affected. In allcontrol groups, animals gained weight, as shown in Table 1.As expected, after 3 days of HSU, animal weights decreasedslightly. After 7 and 14 days of HSU, we observed a slightgain in weight, although not significant from the initialvalues. Food intake (data not shown) was not significantlydifferent among the groups in general, only decreasingslightly during first 2 days of HSU.

As a result of HSU, soleus atrophied significantly (30%,P < 0.05) after only 3 days of HSU (Fig. 1A and B). Thelargest change in the relative soleus muscle mass wasobserved after 7 and 14 days of HSU (47 and 55%,respectively, P < 0.001). The EDL muscle was not affectedby HSU (Fig. 1C and D). However, compared with controlanimals, growth-related increase in EDL mass declined(from 10 to –3%). We found no significant changes in therelative and absolute masses of the gastrocnemius muscleafter 3 days of HSU (Fig. 1E and F); however, GC atrophiedsignificantly after 7 and 14 days of suspension (15 and17%, respectively, P < 0.05). We observed no significantchanges in the cardiac muscle mass (data not shown).

To study gluconeogenesis, animals were subjected toone of the following two fasting protocols: (i) short-term(8 h); or (ii) long term (13 h). To assess the changes in

C© 2012 The Authors. Experimental Physiology C© 2012 The Physiological Society

Exp Physiol 00.00 (2012) pp 1–11 Gluconeogenesis during hindlimb suspension 5

Figure 1. Time course of changes in muscle weights during hindlimb suspension unloading (HSU)A, C and E show absolute weights; B, D and F show weights normalized to body weight. A and B, soleus.C and D, extensor digitorum longus (EDL). E and F, gastrocnemius. The data at 14 days are from bothstudies combined. Data are means ± SEM (n = 4). Significance was determined using Student’s unpairedt test (∗P < 0.05, ∗∗P < 0.001).

C© 2012 The Authors. Experimental Physiology C© 2012 The Physiological Society

6 I. R. Bederman and others Exp Physiol 00.00 (2012) pp 1–11

Table 2. Concentrations of plasma metabolites

Day 3 Day 7 Day 14

Plasma metabolite Control HSU Control HSU Control HSU

Glucose (mM) 9.4 ± 0.712 9.8 ± 0.243 9.7 ± 0.121 11.2 ± 0.245∗ 10.0 ± 0.601 13.5 ± 0.822∗Lactate (mM) 1.5 ± 0.223 1.0 ± 0.225 1.9 ± 0.238 1.4 ± 0.234 1.2 ± 0.212 1.3 ± 0.156Glycerol (mM) 0.10 ± 0.011 0.15 ± 0.001∗ 0.10 ± 0.01 0.14 ± 0.007∗ 0.10 ± 0.013 0.10 ± 0.003Alanine (mM) 0.60 ± 0.072 0.44 ± 0.033 0.53 ± 0.062 0.50 ± 0.052 0.54 ± 0.064 0.61 ± 0.013

Data obtained on day 14 are from both studies combined. Data are means ± SEM (n = 4). Significance was determined using Student’sunpaired t test (∗P < 0.05 compared with control group).

circulating metabolites relevant to fasting, we measuredconcentrations (millimolar) of key plasma metabolites ofcontrol and HSU animals (Table 2). The plasma glucoseconcentration remained normal after 3 days of HSU,but increased after 7 and 14 days. The plasma lactateconcentration was within the expected range. Plasma freeglycerol increased significantly after 3 and 7 days of HSU,but not after 14 days. The plasma alanine (millimolar),released by the muscle for hepatic gluconeogenesis, wasnot significantly different.

The HMT method assumes that C2 2H enrichment isequal to that of body water. Indeed, our measurementsstrongly support the suggestion that there is a fullequilibration of 2H labelling of body water with 2Hlabelling of C2 of glucose (Fig. 2A). 2H labelling of bodywater (control group 2.9 ± 0.02%, versus HSU group,3.0 ± 0.05%) and C2 (control group 2.9 ± 0.06%, versusHSU group 2.9 ± 0.09%) were nearly identical. Thus, itfollows that body water/C2 ratios (Fig. 2 B) would be 1 inall cases (0.97 ± 0.2 for control and 0.95 ± 0.3 for HSU,across all time points).

As the main assumption has been validated, next wepresent findings of fractional hepatic gluconeogenesis ratedetermined using the HMT method (Fig. 3A). There were

no differences among control groups across all timesof exposure to HSU. After 3 days of HSU, GNG wassignificantly higher in HSU animals (75.0 ± 1.5 versus88.4 ± 2.2%, P < 0.01, control group versus HSU group,respectively). However, GNG was only slightly elevatedafter 7 days of HSU (65.4 ± 2.4 versus 71.7 ± 2.2%,P = 0.09, control group versus HSU group, respectively).After 14 days, GNG rate was elevated in HSU animals,but did not reach statistical significance (71.1 ± 4.8 versus82.3 ± 3.2%, P = 0.09, control group versus HSU group,respectively).

To compare analytical methodologies, we determinedthe rate of gluconeogenesis using the average method(Fig. 3B). Similar to Fig. 3A, we found no significantdifferences among the control animals. However, theaverage method yielded significantly lower values thanthose determined using the HMT method. The trend inthe data with respect to changes of GNG in HSU animalswas similar to the findings shown in Fig. 3A; after 3 daysof HSU, hepatic GNG was significantly higher in HSUanimals (47.9 ± 2.0 versus 60.8 ± 2.3%, P < 0.05, controlgroup versus HSU group, respectively), while GNG wasonly slightly higher after 7 (44.6 ± 3.1 versus 49.7 ± 2.6%,n.s., control group versus HSU group, respectively) and

Figure 2. Time course of 2H-labelling during short-term fastingA, 2H labelling of body water (BW) and carbon 2 (C2) of plasma glucose. B, BW/C2 ratio. Data aremeans ± SEM (n = 4).

C© 2012 The Authors. Experimental Physiology C© 2012 The Physiological Society

Exp Physiol 00.00 (2012) pp 1–11 Gluconeogenesis during hindlimb suspension 7

14 days (48.9 ± 5.7 versus 55.3 ± 4.7%, n.s., control groupversus HSU group, respectively). When results from theHMT and average methods were compared, the datadiffered by approximately the same, ∼30% (33.1 ± 1.3versus 31.6 ± 0.5%, for control group and HSU group,respectively). In addition, the GNG rate determined bythe average method was not as significant as determinedusing the HMT method on days 7 and 14.

Finally, we compared analytical methods duringconditions of longer fasting, where gluconeogenesiscontributed 100% to endogenous glucose production. Inthis case, we did not measure a time course of GNG.Instead, we measured the GNG rate after a 13 h fastand after 14 days of HSU (Fig. 4). We first evaluatedthe completeness of equilibration of 2H labelling ofbody water with C2 of glucose (Fig. 4A). Values were inexcellent agreement in both cases (body water 2H labelling,2.86 ± 0.07 versus 2.86 ± 0.06%, control group versusHSU group, respectively; and C2 2H labelling: 2.92 ± 0.08versus 2.92 ± 0.11%, control group versus HSU group,respectively). Next, we compared gluconeogenic ratesdetermined by the HMT and average methods (Fig. 4B).After 13 h of fasting, fractional gluconeogenesis asdetermined using the HMT method was not differentfrom the expected 100% (100.4 ± 0.4 versus 98.7 ± 1.7%,control group versus HSU group, respectively). However,fractional gluconeogenesis, determined by the averagemethod, was significantly lower compared with the HMTmethod for both the control and the HSU group. Inaddition, GNG rates were significantly lower in HSUanimals (70.5 ± 0.3 versus 55.1 ± 2.0%, P < 0.05, controlgroup versus HSU group, respectively). Thus, overall,the average method consistently yielded lower values offractional gluconeogenesis irrespective of conditions offasting and/or HSU.

Discussion

Our goal was to quantify time-dependent changes inthe rate of fractional hepatic gluconeogenesis duringconditions of HSU. Using incorporation of deuterium intonewly synthesized glucose from 2H-enriched body water,we determined the rates of hepatic gluconeogenesis. Wealso compared two experimental methods of analysing 2Hlabelling of plasma glucose, i.e. the HMT method and theaverage method (Landau et al. 1995b; Chandramouli et al.1997; Chacko et al. 2008a). We used two modes of fastingto stimulate gluconeogenesis, short-term and long-termfasting. In the former case, we expected to find relativedifferences between the control animals and suspendedanimals, while the long-term fasting was done to compareanalytical methods in absolute conditions of GNG, i.e.where the rate of gluconeogenesis was expected to be100%. In accord with the hypothesis of Stein et al. (2005)of increased glucogenic capacity of the liver, both plasmaglucose and gluconeogenesis rates were indeed increased.

The hindlimb suspension unloading is a ‘ground-based’technique for simulating ‘zero gravity’. As a result ofHSU, type I fibres of skeletal (‘anti-gravity’) muscles arepreferentially affected, thus causing loss of muscle massand function, i.e. muscular atrophy. Adaptation to HSUoccurs in the first few days (Morey-Holton et al. 2005). Asevident from our data, the animals lost weight after 3 daysof HSU and then slowly began to regain the weight. As thefinal weights of HSU animals were different, we presentdata on changes of the skeletal muscle mass in both relativeand absolute terms. Hindlimb suspension unloadinginduced drastic muscle loss in the soleus muscle, whichis predominantly composed of type I fibres (in Sprague–Dawley rats, ∼83% type I fibres; see Table 1 by Soukupet al. 2002). After 3 days, we observed significant atrophy,

Figure 3. Comparison of the time course of changes in fractional gluconeogenesis during HSU aftershort-term fastingA, fractional gluconeogenesis determined using the Landau method. B, fractional gluconeogenesisdetermined using the Chacko method. Data are means ± SEM (n = 4). Significance was determinedusing Student’s unpaired t test (∗P < 0.05).

C© 2012 The Authors. Experimental Physiology C© 2012 The Physiological Society

8 I. R. Bederman and others Exp Physiol 00.00 (2012) pp 1–11

suggesting exquisite sensitivity of the skeletal muscle to thelack of mechanical stimulation, decrease in energy demandand stress in general. Atrophy worsened as the time ofsuspension progressed; similar data have been presentedby many investigators (Tischler et al. 1985; Thomasonet al. 1987; Wronski & Morey-Holton, 1987; Thomason &Booth, 1990; Taillandier et al. 1996; Bajotto & Shimomura,2006). Also, consistent with observations of others (seereview by Thomason & Booth, 1990), EDL muscle did notchange significantly, even after 2 weeks of HSU. This wasnot surprising, given that EDL consists predominantly oftype II fibres (∼95%; see Table 2 in the review by Soukupet al. 2002). We also found significant atrophy of thegastrocnemius muscle (approximately equal distributionof type I and II fibres). The above findings generallyagree with literature on hindlimb suspension unloading.However, most studies typically show atrophy at only onetime point. In contrast, we measured progressive atrophyover several time points.

To demonstrate the changes in metabolism associatedwith gluconeogenesis, we measured concentrations of keygluconeogenic substrates (lactate, glycerol and alanine).Alanine and lactate concentrations were not altered,whereas glycerol levels were significantly elevated in thesuspended animals after 3 and 7 days of HSU, suggestingelevated rates of adipose tissue lipolysis. This is likelyto be mediated by the well-documented increase incirculating catecholamines and corticosterone (Steffen &Musacchia, 1987; Stump et al. 1992) as the consequence ofunloading. As HSU induces transient adrenal hypertrophy,the increase in hormone levels would be expected. Steffenand co-workers and others found that on day 7 of HSU,corticosterone levels return to normal levels and typicalcyclic pattern (Steffen & Musacchia, 1987; Jaspers et al.1989). Plasma glucose was significantly elevated after 7 and

14 days of unloading. The reasons for this are not entirelyclear. It is possible that unloaded rats had higher ratesof glycogenolysis, stimulated by elevated catecholamines,at 7 day of HSU (Chu et al. 1996), thus raising plasmaglucose. By 2 weeks of unloading, catecholamine levelsshould have returned to normal, but plasma glucose wasstill significantly higher in unloaded animals. This may bedue to the skeletal muscle insulin resistance induced byHSU, which would lead to decreased glucose uptake. Forexample, bed rest was shown to induce insulin resistancein human subjects (Stuart et al. 1988).

After an 8 h fast, the largest increase in thegluconeogenic rate occurred during the acute phase ofadaptation to HSU, i.e. after 3 days. This is likely to bedriven by the increased demand of skeletal muscle forglucose (Henriksen et al. 1986; Henriksen & Tischler,1988a,b; O’Keefe et al. 2004) and in part by elevatedcatecholamines during the initial phase of adaptationto HSU (Thomason & Booth, 1990). By day 7, thegluconeogenic rate decreased transiently as the animalsadjusted to HSU, and by day 14 of suspension thegluconeogenic rates tended to increase but did not reachsignificance. These findings provide novel informationabout adaptation to hindlimb suspension unloading.For the first time, direct measurements of fractionalgluconeogenesis are reported during the time course ofunloading. Our findings confirm previous work of Steinand co-workers (Stein et al. 2005; Stein & Wade, 2005),who demonstrated altered function of the liver, i.e. ametabolic switch towards higher glycogenic function.This is supported by increased gluconeogenic precursors,i.e. amino acids from increased skeletal muscle break-down. The increase in gluconeogenesis supports increasedglycolytic activity and glucose demand in the atrophyingmuscle, as shown by Thomason & Booth (1990).

Figure 4. Comparison of the time course of changes in fractional gluconeogenesis during HSU afterlong-term fastingA, 2H labelling of body water (BW) and carbon 2 (C2) of plasma glucose. B, comparison of hepaticgluconeogenesis as determined by both methods. Data are means ± SEM (n = 4). Significance wasdetermined using Student’s unpaired t test (∗P < 0.05).

C© 2012 The Authors. Experimental Physiology C© 2012 The Physiological Society

Exp Physiol 00.00 (2012) pp 1–11 Gluconeogenesis during hindlimb suspension 9

The rate of fractional hepatic gluconeogenesis wasdetermined using two methods, the HMT method andthe average method. Both methods generated consistentdata that accurately described the HSU-induced increase–decrease–increase trend in the rate of gluconeogenesis.In the original publication of the average method,where Chacko and co-workers also compared the newmethodology with the HMT method, they also showedconsistent results (Chacko et al. 2008a,b). However,the average method yielded lower absolute values ofgluconeogenesis during the conditions of short-termfasting (∼30%) than the HMT method. In addition,during conditions of long-term fasting, i.e. when the GNGrate was expected to be 100%, the average method alsoyielded lower results (∼30%) compared with the HMTmethod. The latter finding agreed with Chacko et al.(2008a), where the rate of fractional gluconeogenesis inhumans fasted for 66 h was 83%, as determined usingboth the average method and the HMT method. Thiswas a surprising finding, because it is well establishedthat after a prolonged fast, hepatic GNG contributes100% to endogenous glucose production (Ekberg et al.1999). Finally, after a 13 h fast, as determined by theHMT method, both WT and HSU animals had 100%GNG, but as determined by the average method the HSUanimals had significantly lower GNG rates (Fig. 4B). Thisindicates not only that the average method underestimatedgluconeogenesis, but also that the measurements weredependent on sample content (see further discussionbelow).

Thus far, systematic discrepancies between the HMTand average methods have been presented, and wereespecially apparent in conditions where GNG is expectedto be 100%. Both methods are based on the precursor–product relationship. As the precursor enrichment, i.e. 2H-labelling of body water, was the same in both methods, theobserved differences can only come from the differencesin product 2H labelling, i.e. 2H labelling of plasma glucose.Indeed, this is where the methods diverge. In the case ofthe HMT method, C5 enrichment was shown to reflectspecifically hepatic gluconeogenesis (Chandramouli et al.1997; Schumann et al. 2001). In the case of the averagemethod, product enrichment is assumed equal for all ofthe glucose carbons and is therefore averaged by dividingby six. Burgess et al. (2008) suggested that this assumptionmay be erroneous, given the complexity of carbohydratepathways; for example, they state that enrichment ofC1 > C5 > C6. Burgess and co-workers provided a strongbiological basis for the enrichment heterogeneity presentin the literature (Leadlay et al. 1976; Chandramouli et al.1997, 1999). Chacko and colleagues, in their defense,stated that the heterogeneity of enrichment of variousglucose carbons is negligible and has little bearing onthe outcome (Chacko et al. 2008b). We demonstrate thatthe average method consistently underestimates GNG,

assuming that the HMT method accurately estimatesGNG. This assumption is reasonable, because we obtainedrates of GNG of the expected 100% after 13 h fasting.Let us consider several possibilities that may result inthe unreliable measurements of 2H labelling of plasmaglucose.

As the main premise of the average method is theuse of the fragment m/z 169, let us first considermass spectrometric issues with this fragment in theconditions of chemical ionization. The main assumptionof the average method is that fragment m/z 169 carriesdeuteriums/hydrogens at C1, C3, C4, C5 and C6, andthose are equally labelled. However, even if one ignoresisomerism from the location of the double bond, thereexist several isobaric structures for m/z 169. Acetate groupscan be found on each of the C2, C3, C4 or C6 positions ofthe carbohydrate ring (Biemann et al. 1963; Guevremontet al. 1990). Wright and colleagues showed in his isotopelabelling experiments (Guevremont et al. 1990) that 16different isobaric species are represented by the m/z 169,and about 30% of these ions were formed via hydrogenshift from acetate to the carbohydrate ring. Competitivefragmentation pathways that lead to the formation ofm/z 169 occur also in accordance with other studies thathave investigated a dissociation behaviour of peracetylatedcarbohydrates (Biemann et al. 1963; Kulkarni et al. 1985).The existence of parallel fragmentation modes, acetaterearrangements and hydrogen transfer from acetate tothe carbohydrate backbone all lead to the formationof m/z 169 isobaric ions and partial loss of deuteriumfrom the carbohydrate ring if the glucose backbonecarries deuterium. This will interfere with enrichmentcalculations that are based on the 169/170 ratio and lead tothe underestimation of m/z 170 by dilution of deuteriumlabelling.

A second consideration is based on several studiesdemonstrating that glucose peracetate has an identicalfragmentation pattern to other hexoses, such as mannoseand galactose (Biemann et al. 1963; Guevremont et al.1990; Cuyckens et al. 2002; Denekamp & Sandlers, 2005) indifferent ionization modes. In particular, all peracetylatedhexoses have fragment m/z 169. Thus, the presence ofmannose and/or galactose will contribute to the measuredfragment of m/z 169 and dilute enrichment. Generally,galactose and mannose are found in small amounts inplasma (1–2%; Schadewaldt et al. 2000; Jozwik et al. 2007).However, if plasma contained measurable amounts ofmannose and/or galactose and those were not properlyseparated via gas chromatography, their presence willcontribute to the formation of fragment m/z 169 and willlead to the overestimation of m/z 169 and consequentlyto the lower enrichment. Such interference can easily beavoided if sugars from the plasma extract are properlyseparated using gas chromatography. Given the inadequategas chromatographic conditions (low initial temperature

C© 2012 The Authors. Experimental Physiology C© 2012 The Physiological Society

10 I. R. Bederman and others Exp Physiol 00.00 (2012) pp 1–11

of 70◦C and fast oven ramp of 30◦C min−1) used inthe average method, it is feasible that the inefficientseparation of various sugars and other compounds foundin a complex mixture such as plasma would interfere withfragment m/z 169 derived from the molecule of glucose.Thus, fragment m/z 169 will represent ion fragments thatoriginated from other sugars and molecules that may ormay not bear deuterium from deuterated water (m/z 170)and thus underestimate labelling to an unknown degree.In addition, the higher plasma glucose concentrationdetected in rats after 14 days of HSU is likely to beresponsible for increasing the interference by unlabelledglucose and causing lower deuterium enrichment, in turncausing a lower GNG rate as shown in Fig. 4B. Theseresults indicate that the average method is sensitive toplasma sugar content and concentration, whereas theHMT method is not.

In conclusion, we show that HSU induces metabolicchanges in the liver. Our novel finding is the markedincrease in the rate of hepatic gluconeogenesis duringHSU. We believe that the primary role for upregulatedgluconeogenesis is to support increased glucose demandthrough upregulated uptake and utilization by the musclesand other organs affected by HSU. Also, a time-dependentincrease in the gluconeogenic rate after short-termfasting was reported irrespective of the analytical methodused. However, in conditions where gluconeogenesis isexpected to be 100%, the HMT method of Landau andcolleagues remains precise and accurate in estimatingthe fractional rate of hepatic gluconeogenesis followinga bolus of deuterated water. The method of Chacko andcolleagues (the average method), although technicallysimple, requires more validation from both analyticaland biological points of view. In particular, the extentof mass spectrometric issues of deuterium loss throughexchange with hydrogen is in need of further evaluationand validation.

References

Bajotto G & Shimomura Y (2006). Determinants ofdisuse-induced skeletal muscle atrophy: exercise andnutrition countermeasures to prevent protein loss. J Nutr SciVitaminol (Tokyo) 52, 233–247.

Bederman IR, Dufner DA, Alexander JC & Previs SF (2006).Novel application of the “doubly labeled” water method:measuring CO2 production and the tissue-specific dynamicsof lipid and protein in vivo. Am J Physiol Endocrinol Metab290, E1048–E1056.

Biemann K, DeJongh D & Schnoes H (1963). Application ofmass spectrometry to structure problems. XIII. Acetates ofpenthoses and hexoses. Am Chem Soc 85, 1763.

Burgess SC, Chandramouli V, Browning JD, Schumann WC &Previs SF (2008). Complicating factors in the application ofthe “average method” for determining the contribution ofgluconeogenesis. J Appl Physiol 104, 1852–1853.

Chacko SK, Sunehag AL, Sharma S, Sauer PJ & Haymond MW(2008a). Measurement of gluconeogenesis using glucosefragments and mass spectrometry after ingestion ofdeuterium oxide. J Appl Physiol 104, 944–951.

Chacko S, Sunehag A, Sharma S, Sauer P & Haymond MW(2008b). Reply to Burgess, Chandramouli, Browning,Schumann, and Previs. J Appl Physiol 104, 1854–1855.

Chandramouli V, Ekberg K, Schumann WC, Kalhan SC,Wahren J & Landau BR (1997). Quantifying gluconeogenesisduring fasting. Am J Physiol 273, E1209–E1215.

Chandramouli V, Ekberg K, Schumann WC, Wahren J &Landau BR (1999). Origins of the hydrogen bound tocarbon 1 of glucose in fasting: significance ingluconeogenesis quantitation. Am J Physiol 277,E717–E723.

Chu CA, Sindelar DK, Neal DW & Cherrington AD (1996).Direct effects of catecholamines on hepatic glucoseproduction in conscious dog are due to glycogenolysis. Am JPhysiol 271, E127–E137.

Cuyckens F, Shahat AA, Pieters L & Claeys M (2002). Directstereochemical assignment of hexose and pentose residues inflavonoid O-glycosides by fast atom bombardment andelectrospray ionization mass spectrometry. J Mass Spectrom37, 1272–1279.

Denekamp C & Sandlers Y (2005). Anomeric distinction andoxonium ion formation in acetylated glycosides. J MassSpectrom 40, 765–771.

Dufner DA, Bederman IR, Brunengraber DZ, Rachdaoui N,Ismail-Beigi F, Siegfried BA, Kimball SR & Previs SF (2005).Using 2H2O to study the influence of feeding on proteinsynthesis: effect of isotope equilibration in vivo vs. in cellculture. Am J Physiol Endocrinol Metab 288, E1277–E1283.

Ekberg K, Landau BR, Wajngot A, Chandramouli V, Efendic S,Brunengraber H & Wahren J (1999). Contributions bykidney and liver to glucose production in the postabsorptivestate and after 60 h of fasting. Diabetes 48, 292–298.

Guevremont R, Pathak V & Wright JLC (1990). Unimolecularand collision-induced dissociation reactions of peracetylatedmethyl glucopyranoside. Organic Mass Spectrometry 25,101–108.

Henriksen EJ & Tischler ME (1988a). Glucose uptake in ratsoleus: effect of acute unloading and subsequent reloading. JAppl Physiol 64, 1428–1432.

Henriksen EJ & Tischler ME (1988b). Time course of theresponse of carbohydrate metabolism to unloading of thesoleus. Metabolism 37, 201–208.

Henriksen EJ, Tischler ME & Johnson DG (1986). Increasedresponse to insulin of glucose metabolism in the 6-dayunloaded rat soleus muscle. J Biol Chem 261,10707–10712.

Jaspers SR, Henriksen EJ, Satarug S & Tischler ME (1989).Effects of stretching and disuse on amino acids in muscles ofrat hind limbs. Metabolism 38, 303–310.

Jozwik M, Jozwik M, Teng C & Battaglia FC (2007).Concentrations of monosaccharides and their amino andalcohol derivatives in human preovulatory follicular fluid.Mol Hum Reprod 13, 791–796.

Katanik J, McCabe BJ, Brunengraber DZ, Chandramouli V,Nishiyama FJ, Anderson VE & Previs SF (2003). Measuringgluconeogenesis using a low dose of 2H2O: advantage of

C© 2012 The Authors. Experimental Physiology C© 2012 The Physiological Society

Exp Physiol 00.00 (2012) pp 1–11 Gluconeogenesis during hindlimb suspension 11

isotope fractionation during gas chromatography. Am JPhysiol Endocrinol Metab 284, E1043–E1048.

Kulkarni P, Kulkarni VS & Pansare VS (1985). Electron impactand chemical ionization mass spectra of sometetra-acetylated anomeric glycosides. Organic MassSpectrometry 21, 23–27.

Landau BR, Fernandez CA, Previs SF, Ekberg K, ChandramouliV, Wahren J, Kalhan SC & Brunengraber H (1995a). Alimitation in the use of mass isotopomer distributions tomeasure gluconeogenesis in fasting humans. Am J Physiol269, E18–E26.

Landau BR, Wahren J, Chandramouli V, Schumann WC,Ekberg K & Kalhan SC (1995b). Use of 2H2O for estimatingrates of gluconeogenesis. Application to the fasted state. JClin Invest 95, 172–178.

Landau BR, Wahren J, Chandramouli V, Schumann WC,Ekberg K & Kalhan SC (1996). Contributions ofgluconeogenesis to glucose production in the fasted state. JClin Invest 98, 378–385.

Leadlay PF, Albery WJ & Knowles JR (1976). Energetics oftriosephosphate isomerase: deuterium isotope effects in theenzyme-catalyzed reaction. Biochemistry 15, 5617–5620.

McCabe BJ & Previs SF (2004). Using isotope tracers to studymetabolism: application in mouse models. Metab Eng 6,25–35.

Morey-Holton E, Globus RK, Kaplansky A & Durnova G(2005). The hindlimb unloading rat model: literatureoverview, technique update and comparison with spaceflight data. Adv Space Biol Med 10, 7–40.

Musacchia XJ, Steffen JM & Deavers DR (1983). Rat hindlimbmuscle responses to suspension hypokinesia/hypodynamia.Aviat Space Environ Med 54, 1015–1020.

O’Keefe MP, Perez FR, Sloniger JA, Tischler ME & HenriksenEJ (2004). Enhanced insulin action on glucose transport andinsulin signaling in 7-day unweighted rat soleus muscle. JAppl Physiol 97, 63–71.

Park E & Schultz E (1993). A simple hindlimb suspensionapparatus. Aviat Space Environ Med 64, 401–404.

Previs SF & Brunengraber H (1998). Methods for measuringgluconeogenesis in vivo. Curr Opin Clin Nutr Metab Care 1,461–465.

Previs SF, Fernandez CA, Yang D, Soloviev MV, David F &Brunengraber H (1995). Limitations of the mass isotopomerdistribution analysis of glucose to study gluconeogenesis.Substrate cycling between glycerol and triose phosphates inliver. J Biol Chem 270, 19806–19815.

Previs SF, Hallowell PT, Neimanis KD, David F &Brunengraber H (1998). Limitations of the mass isotopomerdistribution analysis of glucose to study gluconeogenesis.Heterogeneity of glucose labeling in incubated hepatocytes. JBiol Chem 273, 16853–16859.

Schadewaldt P, Hammen HW, Loganathan K,Bodner-Leidecker A & Wendel U (2000). Analysis ofconcentration and 13C enrichment of D-galactose in humanplasma. Clin Chem 46, 612–619.

Schumann WC, Gastaldelli A, Chandramouli V, Previs SF,Pettiti M, Ferrannini E & Landau BR (2001). Determination

of the enrichment of the hydrogen bound to carbon 5 ofglucose on 2H2O administration. Anal Biochem 297,195–197.

Soukup T, Zacharova G & Smerdu V (2002). Fibre typecomposition of soleus and extensor digitorum longusmuscles in normal female inbred Lewis rats. Acta Histochem104, 399–405.

Steffen JM & Musacchia XJ (1987). Disuse atrophy, plasmacorticosterone, and muscle glucocorticoid receptor levels.Aviat Space Environ Med 58, 996–1000.

Stein TP, Schluter MD, Galante AT, Soteropoulos P, RamirezM, Bigbee A, Grindeland RE & Wade CE (2005). Effect ofhind limb muscle unloading on liver metabolism of rats. JNutr Biochem 16, 9–16.

Stein TP & Wade CE (2005). Metabolic consequences of muscledisuse atrophy. J Nutr 135, 1824S–1828S.

Stuart CA, Shangraw RE, Prince MJ, Peters EJ & Wolfe RR(1988). Bed-rest-induced insulin resistance occurs primarilyin muscle. Metabolism 37, 802–806.

Stump CS, Balon TW & Tipton CM (1992). Effects of insulinand exercise on rat hindlimb muscles after simulatedmicrogravity. J Appl Physiol 73, 2044–2053.

Taillandier D, Guezennec CY, Patureau-Mirand P, Bigard X,Arnal M & Attaix D (1996). A high protein diet does notimprove protein synthesis in the nonweight-bearing rattibialis anterior muscle. J Nutr 126, 266–272.

Thomason DB & Booth FW (1990). Atrophy of the soleusmuscle by hindlimb unweighting. J Appl Physiol 68,1–12.

Thomason DB, Herrick RE, Surdyka D & Baldwin KM (1987).Time course of soleus muscle myosin expression duringhindlimb suspension and recovery. J Appl Physiol 63,130–137.

Tischler ME, Jaspers SR, Henriksen EJ & Jacob S (1985).Responses of skeletal muscle to unloading–a review.Physiologist 28, S13–S16.

Wronski TJ & Morey-Holton E (1987). Skeletal response tosimulated weightlessness: a comparison of suspensiontechniques. Aviat Space Environ Med 58, 63–68.

Yang D, Diraison F, Beylot M, Brunengraber DZ, Samols MA,Anderson VE & Brunengraber H (1998). Assay of lowdeuterium enrichment of water by isotopic exchange with[U-13C3]acetone and gas chromatography–massspectrometry. Anal Biochem 258, 315–321.

Acknowledgements

This work was inspired by the late Dr Marco EugenioCabrera. We dedicate this and the work to his legacy. Hewas a truly inspirational physiologist and will be greatlymissed. This work was supported by grants sponsored bythe National Aeronautics and Space Administration (NASA),‘Digital Astronaut’ NNJ06HD81G (I.R.B. and M.E.C.), NationalInstitute of Diabetes and Digestive and Kidney Diseases DK-14507 (V.C.) and National Cancer Institute 5U54CA116867(Y.S.).

C© 2012 The Authors. Experimental Physiology C© 2012 The Physiological Society