Rat Insulin Turnover in Vivo *

Zucker lean and obese rats were injected under pentobarbital anesthesia with 125I-labeled insulin; at timed intervals from 30 to 120 sec, blood samples were extracted and used for the estimation of insulin levels by RIA. A group of rats from each series was maintained under a constant infusion of noradrenaline. For each insulin determination, a duplicate blood sample containing the same amount of insulin as that used in the RIA, but without the radioactive label, was used as a blank for insulin measurement. The radioactivity in these tubes was then used for the measurement of insulin label per ml blood. From plasma label decay curves and insulin concentrations, the insulin pool size, half-life, and rate of degradation were calculated. Obese rats had higher insulin levels (2.43 nM) and showed less effect of noradrenaline than their lean counterparts, in which insulin distribution volume shrank with noradrenaline treatment. The half-life of plasma insulin was similar in all groups (range, 226-314 sec). Pool size and overall degradation rates were higher in obese (198 femtokatals) than in lean rats (28 femtokatals). It is postulated that obese rats synthesize and cleave much more insulin than lean controls despite their higher circulating levels of insulin.

The rat es of re moval of insu lin, lab eled with radioactive iod ine, from blood may be used to obtain an appro ximate estima tion of ins ulin turnover.However, this approach poses th e problem of di stinguishing between intact ins ulin and other label ed prod uct s of insulin cleavage presu mabl y pr ese n t in plasma .Th e mos t common meth od of m easure ment of insulin concen tra tion in blo od plasma is radioim munoanalysis (27), a pr ocedure that makes use of radioactive iod ine as a label ing ag ent bo und to insul in tyr osinyl residues.Th e use of thi s approac h thwarts an y furth er att em pt to me asure insu lin levels; th rou gh RIA, in a nima ls previou sly injected wi th lab eled insulin .
Th e m ethod p resented here has been dev ised for the s ho rt term es tima tio n of insu lin turnover ill vivo by mea suring th e rate of disap p earan ce of ra d ioiodin e-labeled in sulin in jected into th e blood s trea m co mbined w ith a st a nda rd RIA procedure.
Zucker lean (Fa/?) and obese (fa / fa) male ad ult ra ts, weighin g 324 :!: from Ch a rle s River (Wilmi ngton, MA), were used.The a nima ls were hou sed in individ ua l po lypropyle ne-bo ttomed cages un d er sta nd ard cond itions (lights on from 0800-2000 h; 22-23 C; 70 -75% relat ive humid ity) a nd we re fed s ta ndard chow pe llets (type A04 from Pa nla b, Barcelo na, Spai n).A se ries of six rat s was ca n nu la ted , under 50 mg /kg BW ip pentoba rb ital a nes thes ia) in th e left ca ro tid ar tery (bri ng ing the tip of the ca n nu la just to the heart) a nd in the righ t jug ula r ve in w ith P50 (Clay-Ad a ms, Parsippan y, NJ) pol yethylen e tubes (id , 0.58; od , 0.97 mm), A t the end of th e exp e rim ent, the a ne s the tize d rats were killed by exsan guinati on .Thi s s tudy was co nd ucted in accordance w ith the eu ro pea n corrununity principles, gu ideline s, and proced ur es for a nima l experimentati on .The lean a nd o bese groups of rat s were fur ther s ubdivi d ed in to two g ro u ps; the first was used as control.The an imals of the second group recei ved a cons tant iv in fusio n of 280 ILM noradrenaline (Sigma) th rough a cannula inserted in the low er cava at a ra te of 4 m lz h-kg wh ile un d er pen tobarbital anes thesia .Th e infusion was performed w ith a syringe p um p and maintai ned for 10 min.The rats und er the effects of nor a drenaline infusion w ere used for the es tima tion of ins ulin turnove r, as were their untreated controls.
All work on anesthetized ani mals was carried ou t in a cham ber kept at 32 C and more than 95% rel ati ve hum idity to avoid the h ypothermi c effects of anesthesia.
The resistance to a glucose load was in vest igat ed in five lea n a nd five obese rats of the same age and weight as thos e us ed in the main ex periment.The carotid arteries of the se rats were chronicall y cann ulated under ether anesthesia using P50 p olyethylen e cannulas (29).Tw o d ays after the surgical procedure, the rat s we re given an ora l load of 1 m l glucose (0.7 g) in water via a plastic stomach can nul a.Samples of ca ro tid blood (0.4 ml for lean and 0.2 ml for obese animals) w ere tak en jus t bef ore and 10, 15, and 30 min after the loading of glucos e (0.2 ml for lean and 0.1 ml for obese rats).Thi s blood was used for the me asu remen t of glucose (30) and plasma insulin (31).Figure 1 s ho ws the plasma level s of both.

Insulin turnover measurement
Each rat wa s in jected (in 5-8 sec) through the jugular venous can nula with 18.3 kBq (330 frnol) lab eled ins ulin in 0.1 ml iso tonic saline so lu tio n.The rad ioactivity in itially pr ese nt (an d tha t left a fter the inject ion) in the sy rin ge w as mea sured w ith a v-cou nter.At timed in terva ls of 30, 60, 90, a nd 120 sec, a liq uo ts of 0.4 ml blood we re extr acted throu gh the ca ro tid cannula a nd s to red at 4 C in hep arinized pla sti c via ls.Th e blood sam p les w er e immed iat ely cen trifuged in the cold to se pa ra te the p lasma sam p les, which were used d ire ctl y for lab eled insulin es timati on.Plasma insulin level s w ere est imated by a s ta ndard RIA p rocedure (31) w ith so me minor modifications.Each blood p las ma sa mple was di stributed in tw o tubes; in the first , in a d di tion to the p lasma sample (50-100 ILl) a nd 0-50 ILl buffer, 100 ILl 125I-Iab eled insu lin (74 MBq /nmol; Amersham) so lu tio n con taini ng 250 Bq ti .e. 3.4 frnol) and 100 ILl di lu ted specific insulin antibody (Am ers ha m) were added.In the rem ai nin g tu be, the p lasma samples (50-100 ILl) recei ved 0 -50 ILl buffer , 100 iLl un labeled insu lin (3.4 frnol).and 100 ILl of the specific insulin a n tibody.Th us, the se con d tube finall y contained the same amount of insu lin as the first (the amou nt ini tially p resen t and that added were the sa me as tha t in the first tube), but the amount of labeled insulin present in either w as d ifferent , as the second tube lacked the added labeled insulin used for the standa rd RIA procedure.Because the total amount of insulin was the same in both series of tubes, the labeled insulin initially present in p lasma bo und in the same proportion to the antibody preparation; thu s, the second tube could be used as a blank for the first as in a standard RIA p roc ed ur e.This allowed estimation of the apparent insulin con centra tion (i.e. its binding equivalence to Humulin) regardless of the amo un t and distribution in molecular species of radioactivity initi ally pr esent in the plasma.This approach circumvented the problems po sed by the presence of radioactive sources (insulin and other) in the sam ples.Th e RIA w as completed with a series of standards (of both Hum ulin and ra t in sulin), blanks, and several tubes for the estimation of nonsp ecific binding.
As the concen tra tion of insulin in obese rat blood wa s high er than that in the lea n rat blood, the volume of blood extracted allowed for duplicate d eterm ina tion s of a ll points, whereas lean rats prov ided jus t enough plasma for a sing le determination at each time po int.

Cal culation s
The insulin label pr es ent in a given sa mple of p lasma was esti ma ted assum ing that la be led in sulin w as bound b y th e a n tibo dy in the same propor tio n as u nlabe led ins u lin fro m th e sa me so urce.Rat and huma n in suli n bound th e a nti body differentl y.H ow ever, by using apparen t in sulin co ncen tr a tio ns re ferred to Humulin sta n dard s, th e da ta on ins u lin ra d ioactivi ty co u ld be treat ed as if all ins u lin present in th e ra t bl ood were human in sulin .From th e RIA da ta , a pl ot o f in sulin bound to th e antibod y vs. th e co ncen tra tion of ins u lin in th e tube was d ra wn (Fig .2 progr am (Bioso ft, Cambridge, UK) ; the ca lcu la ted parameters of the curve we re used to es tima te th e pe rcentage of in sulin bou nd to each of the blood sam p les ob tained in the exp er im en t.This percentage also reflected th e p ropo rtion of in sulin rad ioac tiv it y bound to the antibody; thus, the total a m ou n t of insulin ra d ioac tivity (r.) p er ml blood a t a give n time (t) was es ta blis he d for each sa m p le.The r, values were plotted ag ainst time t a nd fitted to a stand ard deca y gr a ph u sing the FiG-P progr am: r, = r o X e-K d ' , from w h ich, K, the decay co ns ta n t, and r o , the initial radioa ctivity pe r ml blo od , were o btain ed .The half-life (t'h) of in sulin w as ca lcu la te d as t'/2 = 11 K .The rat io of total radioa ctivity inje cted (R o ) to r o wa s u sed to establish th e volume (V) of distribution o f the inje cted label (i .e. th e v ir tua l or p ract ical ins u lin space): V = Rol rooThe actual content of huma n insulin wa s ca lcu la ted from tot al in sulin radi oactivit y and the specific acti vit y of th e lab eled insulin injected.Hum an insuli n wa s a m axim um of 4-6 % of the total blood in sulin.Because thi s proportion w as ve ry sma ll, the use of rat insulin standards all owed direct estimation of th e rat insulin con centra tio n in th e sa m p les.The ra t in sulin co ncen tra tion in pl asma di d not va ry during the ex p erim en t.Th e concentration vs .tim e gra phs were used to obtain an est im at e of th e initial ( t = 0) in sulin conc en tration (io), which was pr actic ally identical to th e mean of a ll o ther time po in ts .As w e knew both th e virtua l dis tribu tion volume a nd concen tra tio ns, we could derive the whol e ma ss of circu la ting in sulin (Iv) at tim e zero: 1 0 = i o X V. Th e rate of loss of in sulin (rate of degrad ation, 0) from this circulating po ol co uld be deri ved from the decay curve and the ma ss of in sulin : 0 = K X 1 0 , Indeed , bec ause the virtual distribution volum e (V) did not change, th e in sulin mas s a t a given time (I ,) can be es tima ted from the pla sm a co ncen tra tio ns (i.).and d egradation rates for diffe rent times (0,) may be calculated .The values obt ained in all cases we re sim ilar, becau se th e ch anges in insulin conc entration duri ng the 2-min analys is w ere insig nific ant.The loss of rad ioact ivit y from the label ed insulin poo l was stud ied by establishing the total insulin lab el values (R) at a given time from the radioactivity per ml p las ma and th e virtual volume of di stribution : R, = r, X V.
Statistica l compari son bet ween gro ups was established w ith st anda rd analysis of variance p rograms and Student's t test.

Results
Figure 3 shows the lack of change in plasma insulin levels of the four groups studied during the 2-min analysis of insulin turnover.Zucker obese (fa /fa) rats showed higher insulin levels than lean controls.The infusion of noradren aline slightly decreased obese rat insulin levels, but in lean animals it did not affect insulin levels.
Figure 4 presents the decay curves for plasma insulin radioactiv ity vs .time.Lean and obese rats showed a similar pattern over time.Noradrenaline treatment, however, re sulted in higher radioactivity settings, mainly for lean rats and less for obese, compared with controls.These dec ay curves were used for the calculation of insulin space, turn overrates and cleavage, shown in Table 1.Insulin levels were higher in obese than in the lean rats.Noradrenaline infusion resulted in no significant ch ange in the insulin pool size and a decrease in virtual insulin space.Obese rats had a higher insulin mass than lean controls.The half-life of insulin was in the same range for all groups and was unaffected by noradrenaline treatment.Insulin degradation rates were higher in obe se than in lean rats and were practically un changed by noradrenaline tre atment.These results were maintained even when the data were expressed per unit of animal wei ght.
Figure 5 depicts the fall in total insulin radioactivity in the plasma pool calculated from the data in Fig. 4 and Table 1.The decay curves thus calcu la ted were fully in line with the obese rats ; • and _, noradrenaline (NA )-treated rats; 0 and 0 , un treated controls .The data are the me an z; SE of five or six different r ats per group.Statistical differences between groups were deter mined by two -wa y analysis of variance.There were sign ifica n t dif feren ces between lean and obese gr oups (P = 0.000).The effects of noradrenaline treatment were s ign ifica nt (P = 0.039) for obese, but not for lean, rats.Th e effects of time were n ot si gnificant C P > 0.05) in any ca se.
zero tim e values, corresponding to the initial (injected ) la beled insulin mass.Lean rats showed a higher loss of abso lute insulin radioactivity than the obese; under noradrena line infusion, the differences between lean and obese rats were minimal in absolu te terms.

Discussion
The method presented is conceptually very simple, but requires careful development, especially at the critical point of evaluation of insulin radioactivity in the samples.This is dependent on precise measurements of radioactivity and the construction of a well defined RIA st andard curve, from which an equation could be derived.Another critical point is the need for high insulin specific activity; this is essential for the precise RIA estima tion of circulating insulin, but it is equally preferable for injection into the rat bloodstream, be cause the higher the specific activity, the lower the distur bance.In our case, the insulin injected was about 7% of the whole circulating insulin pool in lean controls and 0.6-1 % of that in obese rats.
A critic al point in the investigation of insulin cleavage is the ass ignment of radioactivity measurements to intact (i .e. fully functional) insulin, without interference by free iodine or labeled fra gments freed by the cleavage of insulin.The method presented prevents this interference, as onl y the label bound to insulin is measured; free iodine and labeled pep tides are removed during the RIA procedure.Only labeled, complete, insulin is bound to the a ntibody, and thus onl y this

RAT INSULIN TURNOVER IN VIVO
Vol 136 • No 9 FIG. 4. Insulin r adio activity decay with time in the plasm a of Zu cker lean a nd obese rats injected with labeled in sulin, Data represent th e m ean (:!:SE) r adioactivity in 1 ml plas m a a s a per cen tage of the total r a dioactivi t y in j ect ed , found at given t im es a fter inj ec tion .Th e cu r ves we re fit to ex ponen tia l de cay cur ves with r values of 0.989 a nd 0.978 for lean a nd obese cont rols, re spect ively, and 0.994 a nd 0.979 for lea n and obese rats during nor adrenaline infusion.radioactive molecular species is taken into account; this is true for both the insulin injected into the rat and that used in the RIA.
The main difficulty that may arise from a study based on calculated constants taken from calculated values and used to derive the final results is a cumulative effect of residuals in calculations that m ay lead to widening errors.This study has been designed to minimize this effect.The ad equacy of decay curve fitting is apparent in Fig. 3, which shows the loss of radioactive insulin per ml plasma.The d ata for total in sulin radioactivity remaining in the rat we re calcula ted by applying the data derived from these equations and insulin levels.Figure 4 shows the tight fit of th ese calculated data to de cay curve equations.The lack of dispersion of data sug gests an acceptable degree of precision in the derivation of the virtual volume of d iffusion and decay rat es shown in Table 1.
The compartments occupied by the circulating insulin pool are probably not uniform with respect to insulin con centration (22,24), as th is is affected by the varying ability of tissues to extract an d inactivate insulin (12,13).It is difficult to establish the me an representative values of circulating insulin in a given subject, because insulin levels may differ substantially in different blood vessels and under distinct cond itions and times .For this reason, only blood from the same sou rce wa s an alyzed.The carotid cannula allow s the extr action of arterial blood just as it lea ves the heart.The virtual volume of d istr ibution of insulin derived from data of arterial blood may not be real, because we do not know whether its insulin levels are representative.The insulin pool size (I t) is more reliable, because it may be de rived directly from decay curves and insulin levels.Virtual insulin space w as higher in lean th an in obese rats under standard conditions; in an y case , thi s value was higher than the total rat plasma vo lu me, but mu ch sm aller than the sum of insulin compartments (24).Obese rats ha ve a lower extracellular space than lean because of large fat deposits, Insulin pool size was m uch higher in obese than in lean rats, both in absolute terms and in relation to body weight; in the latter case, the differences were somewhat diminished bl ' " cause of the dilution effect of the large fat mass in obese rats Hyperinsulinemia is asso ciated with obesity (32), as iti: related to higher fat deposition and insulin resistance (33).The apparent chan ges obs er ved in lean rats as a consequence of catecholamine infusion may be due in part to the shrinkage of insulin virtual space, largely the space occupied by the blo od , i.e. blood vessels, because of vasoconstriction (34,35 1  The significan t changes in pl asma insulin radioactivity decal  5. Total ra dioactivity present in the plasma insulin pool of Zucker lea n and obese rats after th e inj ection of la beled insulin .The da ta are the mean ( :!:SE) of t he quotien ts bet ween rad ioac tivity fouo d (t hat in 1 ml plasma m u lti plied by the virtual volume of distribu tion) a nd th at injected for the time given.may be the consequence of a low er d istrib u tio n space; the size of the insulin pool changed less tha n the insulin levels.However, in obese rats the da ta pr esented show a lack of respo nse of insu lin poo l d y na mics to noradrenaline in fu sion; the e ffects of noradrenaline were less ap pa ren t than those in lea n ra ts, cha n ging the ins u lin s pace and degra dation rate only fractionall y.This lack of effect may be related to higher basal ad renergic stimulation in obese rats related to hyp er tension (36) and co un te ra ctive adaptation to hyperinsulinemia (37,38).
The obese rats used in this experimen t showed a marked resistance to glucose, with hyperinsulinemia and basal nor moglycemia, as described for Zucker fa /fa rats (37).
Insulin ha lf-life was sim ilar in lean and obese rats; all values we re in the 4-min ran ge .The similarities be tween groups in this res pec t were, however, on ly apparen t, as the estimated ra tes of ins ulin inactiva tion were much higher in obese than in lean rat s ( ~ 7-fold under the conditions tested ).Noradrena line injection had no significan t effect on these rates in eit her gro up of ra ts.
There is a direct relat ion sh ip in hu mans bet ween fat ac cumulation and built-in capability to inactivat e insulin (39), which is pa rtly confi rmed by the finding that obese rat s cleave insulin much faster than lean controls.The main te nance of higher insulin leve ls with sim ilar ha lf-live s for in sulin mo lecules im pli es a higher ins ulin turnover in the obe se animals.The increased d egradation of insulin with ma inte nance of hig h circula ting lev els and pool size im pli es in creased secretion (synthesis) by the pancreas .It ma y be po s tulated that the cons tan t release of lar ge amounts of ins ulin (in the ran ge of 210 /Lg /d ay'k g in obese rats compared with -38 /Lg / day-kg in lean animals) may be instrumental in the development of lat e-onset diabetes often associated with obesity.

Acknowledgment
Thanks are giv en to Robin Rycrof t for ed itoria l assis tance.

FIG. 1 .
FIG.1.Effect of an oral glucose load on plasma insulin and glucose levels in lean and obese Zucker rats.The data are the m ean z SEM of five different animals per group.The load (0.7 g glucose ) was given a t time zero.

10 TOTAL
) u sing Humulin s ta nda rds .The d atil were fitted to a n as ymmetric s ig moid curve us in g th e FiG-INSULIN [nMI F IG. 2. Rel ation ship between the human insulin concentration found in the RI A wit h th e proportion of labeled insulin bound to the antibody .The d ata wer e fitted t o a n eq uation wit h the formula: y = Min + (Max -Min )/( 1 + [(XlX50)-PJ), wher e X50 = 4.525 x 10 8 ; P = 0.2673 ; Max = 943.017; and Min = -4 .1240(r = 0.992).

FIG. 3 .
FIG.3.Pl a sma insulin concentrations in Zucker lean and obese rats after the injection of labeled in sulin .0 a nd ., Lean r ats; 0 and _ , zer o shows t he un ity, i.e. t he in (f) ~ jected ra dioactivity only.The curves were fittoexpone ntial decay curves, with r val < ~ 0.25 ues of 0.998 and 0.977 for lea n a nd obese (f) cont rols, resp ectively, and 0.994 and

TABLE 1 .
Com pa rison of inj ected labeled in sulin turnover in a nesthe ti zed lean a nd obese Zu cker rats : effect of noradrenaline infusion