CARBOHYDRATE METABOLISM INPREGNANCY. I. THE METABOLISM OF INSULINBY HUMAN PLACENTAL TISSUE

Norbert Freinkel, Charles J. Goodner

J Clin Invest. 1960;39(1):116-131. https://doi.org/10.1172/JCI104010.

Research Article

Find the latest version:

http://jci.me/104010-pdf

CARBOHYDRATEMETABOLISMIN PREGNANCY. I. THE METABOLISMOF INSULIN BY HUMANPLACENTALTISSUE* t

By NORBERTFREINKEL ANDCHARLESJ. GOODNERt(From the Thorndike Memorial Laboratory and Second and Fourth (Harvard) Medical Serv-

ices, Boston City Hospital, and the Department of Medicine, Harvard Medical School,Boston, Mass., and the Howard Hughes Medical Institute)

(Submitted for publication July 22, 1959; accepted August 19, 1959)

Much clinical evidence would suggest that thereis a change in carbohydrate metabolism duringpregnancy (2-8). In the past, the diabetogeniceffects of pregnancy have been attributed to an in-creased elaboration of hormonal contrainsulin fac-tors. An alteration in the degradation of insulinduring pregnancy might constitute a supplementalor alternative mechanism. It has been shown byMirsky, Perisutti and Dixon (9, 10) and Tomi-zawa, Nutley, Narahara and Williams (11, 12)that various animal tissues, such as liver, containrelatively specific enzymes for the proteolytic in-activation of insulin. To date, such systems havenot been studied in isolated preparations of humantissues. In the present study, the presence of anactive, enzymatic mechanism for the proteolyticdegradation of insulin has been demonstrated inhuman placental tissues. The characteristics ofthe system have been defined and its potential con-tributions to maternal insulin economy have beenevaluated.

MATERIALS

1. I1"'-insulin. Crystalline beef insulin (Lilly zinc in-sulin No. 535664 assaying 27 units per mg.) was pur-chased from Abbott Laboratories, Oak Ridge, Tennessee,in solutions containing 0.1 mg. insulin per ml. It hadbeen labeled with I'31 to an initial specific activity of 4to 7 mc. per mg. The iodination entailed the substitution"approximately one atom of iodine per molecule of in-sulin, assuming a molecular weight of 6,000 for the in-sulin" (13). Upon receipt in this laboratory, the I1`-insulin was immediately diluted with four parts of phos-phate-buffered sodium chloride (116 mMoles per L.)

* This investigation was supported in part by ResearchGrant A-1571 (C) and Training Grant 2A-5060 (C2),National Institute of Arthritis and Metabolic Diseases,United States Public Health Service, Bethesda, Md.

t Presented in part at the joint meetings of the Ameri-can Society for Clinical Investigation and AmericanFederation for Clinical Research, May 4, 1958, AtlanticCity, N. J. (1).

t Public Health Service Research Fellow of the Na-tional Institute for Arthritis and Metabolic Diseases.

containing 5 per cent human serum albumin (HSA).Thereafter, the material was dialyzed at 40 C. for 24hours against serial changes of sodium chloride (135mMoles per L.) buffered at pH 7.4 with sodium phos-phate (10 mMoles per L.). The resultant dialyzed andbuffered albumin-insulin mixtures were employed as stocksolutions. No stock solution was employed for morethan 10 days. Between experiments, stock solutions werestored in the frozen state.

More than 98 per cent of the total radioactivity in thestock I"'1-insulin solutions could be precipitated with tri-chloroacetic acid (TCA). Various preparations wereanalyzed for homogeneity by hydrodynamic flow paperchromatography (14) and by paper electrophoresis onWhatman No. 3 filter paper, in barbituric acid-sodiumbarbiturate buffer, pH 8.6, r/2 0.05. As judged by thelatter criteria, products of radiation damage or spontane-ous degradation did not exceed 7 per cent of the total ra-dioactivity. When three separate stock solutions were di-luted to 1 X 0I units per ml., their hypoglycemic po-tency in the rat epididymal fat pad assay system (15)was not significantly different from that of comparablequantities of unlabeled crystalline insulin.

For individual experiments, stock solutions of I`-insulin were supplemented with unlabeled "glucagon-free"crystalline zinc insulin.1 The unlabeled insulin had beendissolved in water with the addition of a few drops of0.01 N HCOand solutions of 2 mg. insulin per ml. weremaintained at pH 3.9 during storage at 4° C.

2. Placentas. No tissues from patients with obstetricalcomplications or disturbances of carbohydrate metabo-lism were included in the present study. Term placentaswere obtained following vaginal delivery or elective re-peat Caesarian section. Retroplacental blood was ex-pressed manually. Immediately thereafter, placental seg-ments were excised with blunt dissection scissors, freedfrom adherent membranes, briefly blotted on filter paper

1 Crystalline zinc insulin (No. 499667) assaying 25units per mg. and crystalline glucagon (No. 258-235-B-54-2) were generously provided by Dr. 0. K. Behrensand Dr. C. W. Pettinga of Eli Lilly Co. Somar-A(Bovine Growth Hormone), Armour No. R50109, andPanlitar (Ovine Lactogenic Hormone), Armour No.R10109, were gifts of the Endocrinology Study Section,National Institutes of Health. Thytropar (Bovine Thy-rotropin Armour No. R2402), Bovine CorticotropinACTH (Armour No. R68406C) and crystalline bovinealbumin (Armour), were obtained commercially.

116

PLACENTALINSULIN DEGRADATION

and frozen on solid carbon dioxide. All manipulationswere completed within 5 minutes following delivery ofthe placenta. Weights of the placentas and membraneswere measured with a single pan balance, accurate to± 1 Gm. For experiments involving placental slices, thetissues were not frozen but rather wrapped in cellophaneand transferred to the laboratory on cracked ice inDewar flasks.

Obstetrical management of the patients consisted ofroutine premedication with 90 to 120 mg. Seconal® and0.4 mg. atropine. Caesarian sections were performedwith spinal anesthesia. In some instances 1 ml. (10 I.U.)Pitocin® was administered intramuscularly prior to ex-pulsion of the placenta. Placental enzymatic activity didnot appear to be significantly affected by these medica-tions. Comparable results were observed in two ex-periments in which placentas were obtained followingprecipitous delivery before any drugs could be adminis-tered.

METHODS

1. Slice experiments. Tissue slices of 0.5 mm. thick-ness were prepared by free hand slicing, floated on chilledsaline, and washed in three portions of saline to removeadherent blood. Two hundred + 10 mg. slices were in-troduced into open 20 ml. beakers containing 2.0 ml. ofmodified Krebs-Ringer phosphate enriched with 2 mg.per ml. of glucose (KRP-Glucose). Following tempera-ture equilibrations, 0.2 ml. of insulin solutions was in-troduced and incubation was performed in the Dubnoffapparatus in an atmosphere of 100 per cent oxygen.

2. Homogenate experiments. Weighed frozen ali-quots of placenta were homogenized in Potter-Elvehjemground-glass homogenizers with two parts 0.1 M, pH 7.4,phosphate buffer to yield 1: 3 homogenates. For thestandard experimental procedure, the homogenates werecentrifuged for 30 minutes in an International refriger-ated centrifuge at 15,000 x G and 40 C. One-half ml.aliquots of the supernatant solutions were added to 1 ml.of 0.1 M, pH 7.4, phosphate buffer contained in open 20ml. reaction vessels. Insulin solutions were prepared bycombining diluted stock solutions of I`31-insulin with vary-ing quantities of unlabeled insulin in 0.1 M, pH 7.4, so-dium phosphate buffer containing 25 mg. of HSA perml. Enzyme and substrate preparations were equilibratedat 380 C. Thereafter, 0.5 ml. quantities of the insulinsolutions were introduced into the reaction vessels andincubation was performed at 380 C. in the Dubnoff appa-ratus (shaking rate, 100 cycles per min.). Thus, the finalreaction mixture in the standard experimental procedureconsisted of 2 ml. of pH 7.4 phosphate buffer containing12.5 mg. of HSA, varying amounts of insulin, and the 15,-000 X G soluble supernatant solution derived from 166 mg.of wet weight placenta. Control studies had indicatedthat the 6.2 mg. HSA per ml. did not alter enzymaticactivity although it prevented adsorption of insulin toglassware and the spontaneous degradation of diluteiodoinsulin (vide infra). Moreover, this amount of

HSA provided sufficient carrier protein for precipitationwith TCA. In all studies, concurrent incubations wereperformed with blank reaction vessels containing onlybuffer and the insulin-HSA mixtures.

3. Assay of insulin degradation. Specific techniquesemployed to characterize the products of insulin degrada-tion will be discussed separately in the Results section.In the standard experimental procedure, assay was basedon the solubility characteristics of the radioactivity inTCA. At timed intervals throughout incubation, 0.2 ml.aliquots of reaction mixture were removed and intro-duced into 13 X 100 mm. counting tubes containing 2 ml.of 20 per cent TCA. Radioactivity was partitioned intoTCA-soluble and TCA-insoluble fractions by centrifuga-tion. The insoluble residues were washed by resuspen-sion in 2 ml. of 20 per cent TCA and centrifuged. TheTCA supernatant fractions were pooled and the washedinsoluble moieties were dissolved in 4 ml. of 30 percent KOH. Radioactive assay of the TCA-soluble andTCA-insoluble fractions was performed in a Nancy Woodwell-type scintillation counter with a sensitivity of 1.0 X10' cpm per jsc. I"f and background of 125 cpm. Suffi-cient counts were observed to reduce the probable errorof the measurement to less than + 3 per cent. After ap-propriate corrections for blanks, the TCA-soluble radio-activity in individual specimens (i.e., nonprecipitable 131)was expressed as a percentage of the total radioactivitywithin the system (i.e., the sum of the TCA-soluble plusTCA-insoluble radioactivity).

In the slice experiments, 2 ml. of 20 per cent TCA wasadded to the vessels at the end of incubation. The flaskcontents were quantitatively transferred to Potter-El-vehjem ground-glass homogenizers and were homogenizedpreparatory to fractionation into TCA-soluble and TCA-insoluble components as above.

4. Analytical techniques. Paper electrophoresis wasperformed with the apparatus and techniques describedelsewhere (16). Butanol-acetic acid-water 78:10:12(B Ac) and butanol-dioxane-2 N NH3 4:1:5 (BDA)were employed for chromatography. Herein, portionsof untreated reaction mixture, or TCA-soluble super-natant were applied to Whatman No. 1 filter paper afterbeing supplemented with carrier L-thyroxine (T4), L-tri-iodothyronine (T3), L-monoiodotyrosine (MIT), L-di-iodotyrosine (DIT), sodium iodide, and 1 X 10' M1-methyl-2 mercaptoimidazole. The carrier compoundswere localized by successively spraying with solutions of0.1 per cent palladium chloride and diazotized sulfanilicacid. Duplicate chromatograms were also sprayed with

*0.1 per cent ninhydrin in acetone. Distribution of radio-activity in individual electrophoretograms and chromato-grams was qualitatively localized by radioautography andquantified with the use of an electronically integrating,automatically recording, strip-scanner developed in thislaboratory (17). A micro-Kjeldahl method was em-

ployed for analysis of total nitrogen in duplicate ali-quots of placental extracts. Blood glucose was meas-ured by the method of Somogyi as modified by Nelson(18).

117

NORBERTFREINKEL AND CHARLESJ. GOODNER

RESULTS

1. MethodsThe published findings of ot]

prompted a re-examination of the pc

dological artifacts in the use of I13vitro systems. Representative ex

described in detail below.A. Adsorption to glassware. Ter

ml. beakers were filled with constartities of I131-insulin, varying amountunlabeled insulin, and sufficient 0.1 Mphate buffer to yield 2.0 ml. syste0.25 to 500.25 ug. insulin per ml. (1vessels were counted by being placaluminum disk directly above a s5tector head. Thereafter, they were

30 minutes at 380 C. in the DubrFollowing incubation, the beakersinto separate 10 ml. volumetric fla50 mg. HSAand refilled with 2.0 mbuffer. Repeat radioactive assay )

to assess the "percentage of the itivity" which had been retained wit}("glassware adsorption"). This encent of the initial counts in the vesse

nally did not contain more than 2.5ml. (Column A, Table I). The ev

TABLE I

Adsorption of insulin to glassware and spotion of insulin during incubation in al

Reaction mixture: 2 ml. pH 7.4 phosphate buffe500.25 pg. insulin per ml. and no other protminutes, 380 C.

XT- T--.I.l T-T A A*co. insunln rbA tA*

pg./mi. mg./ml. %1 0.25 0 46.72 0.75 0 48.73 1.50 0 32.94 2.50 0 27.65 5.00 0 15.36 10.25 0 8.57 50.25 0 4.88 100.25 0 3.59 200.25 0 3.1

10 500.25 0 3.9

* A: "per cent of initial radioactivity"after decanting the original 2.0 ml. su,(i.e., "glassware adsorption").

t B: "per cent of initial radioactivity" iafter washing successively with 2.0 mlphosphate buffer.

jC: minimal estimate of "per cent ofity" rendered soluble in TCA during"spontaneous degradation").

the "adsorbed" radioactivity could be removed bywashing was assessed as follows: beakers were

hers (19-22) decanted into the same volumetric flasks; a singleDtential metho- washing with 3.0 ml. of phosphate buffer was per-.-insulin in

. formed; then the beakers were counted for a third:periments are time. As shown in Column B, Table I, more than

5 per cent of the initial radioactivity still remained

uncovered 20 in these beakers.B. Spontaneous degradation. The albumin-con-

s of crystalline taining 10 ml. volumetric flasks into which the in-

[, pH 7.4, phos- sulin and the beaker washings had been decanted

.mscontaining in the above experiments were employed to as-

Fable I). The sess spontaneous degradation. Control volumetric

ced on a fixed flasks were prepared by directly combining 0.50

cintillation de- Pug. I131-insulin with 50 mg. HSA without pre-! incubated for liminary incubation. Flasks were diluted to con-

ioff apparatus. stant volume and duplicate 1.0 ml. aliquots fromeach were fractionated into TCA-soluble and TCA-were emptied insoluble components. After appropriate correc-

osks containing tion for TCA-soluble radioactivity in control flasks,1. of phosphate minimal estimate of spontaneous degradation dur-wvas performed minitial radioac- ing incubation was obtained by multiplying a) the

in each beaker percentage of the initial radioactivity which had

Kceeded 25 perbeen transferred to the volumetric flasks (i.e., 100

Is which25 per per cent minus Column B, Table I) by b) the per-

,u isuorigi- percentage of the radioactivity in the volumetric flasks

Mg. insulin per which was soluble in TCA. Results are listed inase with which Column C, Table I. During aerobic incubation of

aqueous systems containing less than 2.5 Mg. insu-lin per ml., more than 5 per cent of the insulin is

ntaneous degrada- rendered soluble in TCA.queous m~ined Similar values for glassware adsorption and

containing 0.25 to spontaneous degradation were obtained in repeatlein. Incubation: 30 studies with varying proportions of I131-labeledBt ct and crystalline unlabeled insulin. Thus, the phe-

nomena seem to reflect the fate of all of the insulin9.6 8.9 within the system rather than merely the labeled

13.8 7.9 material. However, some adsorptive peculiarities7.4 640 of the J'31-insulin due to modification of the phe-3.8 3.9 nolic hydroxyl in the iodinated tyrosyl groups2.3 1.21.4 0.7 cannot be wholly excluded.0.7 0 C. Effects of human serum albumin on adsorp-0.6 0 tion to glassware and spontaneous degradation.

Similar protocols were employed to assess theretained in beaker effects of HSAupon the adsorption to glasswarespending medium and spontaneous degradation of dilute insulin solu-retained in beaker tions. For these studies, open 20 ml. beakers were1 and 3.0 ml. of filled with 2.0 ml. of phosphate buffer (pH 7.4)initial radioactiv- containing constant amounts of labeled and un-incubation (i.e., ' labeled insulin (0.25 uig. insulin per ml.) and

118

PLACENTAL INSULIN DEGRADATION

TABLE II

Effect of human serum albumin upon adsorption of insulin toglassware and spontaneous degradation of insulin during

incubation in aqueous media

Reaction mixture: 2.0 ml. pH 7.4 phosphate buffer containing 0.25 pg.insulin and 0 to 10.0 mg. HSA per ml. Incubation: 30 minutes,380 C.No. Insulin HSA A* B* C*

pg./ml. mg./m.1 0.25 0 46.8 21.1 5.32 0.25 0.05 16.3 6.4 2.03 0.25 0.10 8.7 2.3 04 0.25 0.20 7.9 2.1 05 0.25 0.50 6.1 1.6 06 0.25 1.00 5.9 1.6 07 0.25 2.00 6.2 1.4 08 0.25 5.00 4.3 0.8 09 0.25 10.00 4.4 0.7 0

* As in Table I.

varying amounts of HSA (0 to 10.0 mg. per ml.)(Table II). After 30 minutes of incubation ofthese beakers, the adsorption of radioactivity toglassware in vessels containing no HSAaveraged46.8 per cent as above (Table II, Column A).However, with as little as 0.10 mg. HSA per ml.in the suspending media, this adsorption was re-duced to below 10 per cent. Moreover, when theinitial suspending medium contained 0.50 mg.HSA per ml., two aqueous washings sufficed torecover more than 98 per cent of the radioactivity(Table II, Column B). Finally, the spontaneousdegradation of insulin during 30 minutes ofaerobic incubation did not occur in media con-taining 0.10 mg. HSAper ml. (Table II, ColumnC).

Additional studies were performed to evaluatethe reversibility of glassware adsorption. Morethan half of the insulin adsorbed to glass during30 minutes of incubation of aqueous media con-taining 0.25 ,ug. insulin per ml. could be recoveredby introducing 25 mg. HSA into the system for5 minutes prior to decanting.

Experiments to be published elsewhere wouldindicate that HSA does not contain insulin, andthat the effects of HSAare not due to an interac-tion between HSAand insulin. Rather, at pH 7.4,insulin and HSAseem to compete for similar bind-ing sites on the glass surface.

D. Conclusions. Significant adsorption to glass-ware and spontaneous degradation of insulin mayoccur during incubation of dilute insulin solutionsin open vessels. The proportional magnitude of

the phenomena at pH 7.4 is conditioned by a) theconcentration of insulin, b) the presence of othernegatively charged proteins, such as HSA, and c)the relationship between surface area and volumeof medium. The contribution of such methodologi-cal artifacts can be minimized by the inclusion ofHSA in all suspending media.'

2. Degradation of insulin by placental structuhres

Placental structures were prepared in variousways to assess their capacity to degrade insulininto TCA-soluble products.

A. Comparison of slices versus homogenates.Results of one of three experiments are presentedin Table III. Slices of chorion, amnion, and pla-centa (200 10 mg. wet weight), and 0.2 ml. ali-quots of maternal or cord blood were incubated in2 ml. KRP-Glucose containing unlabeled insulinand tracer quantities of I13'-insulin. Cell-freesystems were prepared by homogenizing 1.0 Gm.of slices or 1 ml. of maternal or cord blood with4 ml. of KRP-Glucose and then incubating 1 ml.aliquots of the homogenates with the same amountof insulin in a final volume of 2.2 ml. During 60minutes of incubation, all of the slice systemsliberated appreciable quantities of "nonprecipitableI13113 (Table III). Placental slices were most ac-

tive on the basis of wet weight. However, none

of the slice preparations was as active as homoge-nates prepared from equivalent amounts of tissue

TABLE III

The degradation of insulin by placental structures

Reaction mixture: KRP-Glucose, pH 7.4 plus 200 i 10 mg. cellularpreparation or 1 ml. of 1:5 homogenate. Final insulin concen-tration 25.02 ug. per ml. Incubation: 60 minutes, 380 C. Fordetails, see text.

Nonprecipitable I131(% of total 1131)

Slices Homogenate

Tissue:Chorion 11.5 57.1Amnion 10.6 51.3Placenta 36.6 67.5Maternal blood 1.3 0.1Cord blood 0.5 0.1

2 Many other proteins are equally effective. How-ever, the concentrations of HSA which prevent adsorp-tion to glassware and spontaneous degradation do notaffect enzymatic proteolysis of insulin whereas some ofthe other proteins, such as Cohn Fractions III and IV-1,contain noncompetitive inhibitors of insulin-degradingenzymes (23).

119

NORBERTFREINKEL AND CHARLESJ. GOODNER

I-.k 60

, 404.

20 .

Extractseooted

TOTAL HOMOGENArESOLUBLESUPERNATEafter 30min.;I5,000x

EZJ SOLUBLESUPERNArE after 60min.;I00,000z.

5 min. at 381C 5 min. at 550C 5 min. at 100.C

FIG. 1. INSULIN DEGRADATIONBY SUBCELLULARFRAC-TIONS OF HUMANPLACENTA: THERMAL LABILITY OF

SUBCELLULARFRACTIONSPlacental fractions were prepared as outlined in text

and heated for five minutes at 380, 550 or 1000 C. priorto incubation. Nitrogen contents of the 1: 3 total ho-mogenate, 15,000 X G and 100,000 X G soluble super-

natants were 689.7, 455.6, and 336.4 mg. per 100 ml., re-

spectively. Incubation: 30 minutes, 380 C. Insulin con-

tent of reaction mixture: 50.01 ,tg. per ml.

(Table III). The findings cannot be attributedto occluded blood since negligible degradation was

effected by maternal or cord blood.B. Subcellular localization of placental activity.

A portion of a 1: 3 placental homogenate was

stored at 40 C. while other aliquots were centri-fuged either in a) an International Refrigeratedcentrifuge at 40 C., 15,000 X G for 30 minutes tosediment cellular debris, nuclei and most mito-chondria or in b) a Model L Spinco preparativeultracentrifuge at 40 C., 100,000 x G for 60 min-utes to sediment most of the microsomes in addi-tion to the denser subcellular components. Por-tions of intact homogenate and the two supernatantsolutions were heated for 5 minutes in water bathsat 380, 550 and 1000 C. Thereafter, 0.5 ml. ali-quots of the various heated and unheated prepara-

tions were incubated for 30 minutes at 380 C. with1.5 ml. of insulin-containing phosphate buffer.As shown in Figure 1, most of the insulin-de-grading capacity of human placenta was localizedin the soluble cytoplasm which remained aftercentrifugation at 100,000 x G. Specific activityexpressed as "per cent nonprecipitable I131 per

mg. N of placental extract" was increased 60 to

200 per cent by such centrifugation. The activitywas largely, although not entirely, destroyed by 5minutes of boiling.

C. Conclusions. Humanplacenta contains heat-stable and heat-labile systems for liberating TCA-soluble radioactivity from I131-insulin. Activityis maximal in cell-free systems and localized prin-cipally in the nonparticulate soluble cytoplasm.

3. Products of insulin degradation by placentalextracts

Additional techniques were employed to charac-terize the products of the reaction between in-sulin and placental extracts. For these and forall subsequent studies, placental homogenates wereprepared and incubation was performed accordingto the standard experimental procedure.

A. Paper electrophoresis. In dilute solution,I131-insulin remains bound at the origin during pa-per electrophoresis in veronal buffer at pH 8.6 (14,21, 24). However, paper electrophoresis of thereaction mixture following incubation of I13L-insulin and placental extract disclosed radioactiveproducts which migrated in the areas correspond-ing to the slowest moving portion of albumin andthe a2-globulins (Figure 2). In addition, occa-sionally a small amount of radioactivity was notedin the prealbumin area. More than 90 per cent ofthe applied radioactivity could be accounted forin these areas. Under similar conditions, recoveryof inorganic 1131 averages 3.1 per cent (16).When multiple reaction vessels were preparedwith constant tracer quantities of IJ3"-insulin andvarying amounts of crystalline unlabeled insulin,the formation of radioactive degradation productswas inversely proportional to the total amount cfinsulin within the system (Figure 2). Althoughthe characteristic electrophoretic migration at pH8.6 of the radioactive degradation products mayhave resulted in part from associations with plasmaproteins, a similar migration was exhibited duringelectrophoresis of the supernatant solutions ob-tained following TCA-precipitation of the reactionmixtures.

B. Paper chromatography. In both acid (BAc)and alkaline (BDA) solvents, the radioactivity inthe reaction mixture at all intervals following in-cubation was apportioned between the origin andareas which corresponded to the migration of the

I

120

PLACENTAL INSULIN DEGRADATION

EXTRACT: 15,000 x g. Soluble Supernate

ORIGIN

INCUBATION: 60minutes, 38C*INSULIN

(,ug. f mL.1

12.1

5 .1

0.6

T 1ALBUMIN

FIG. 2. ELECTROPHORETICMIGRATION OF PRODUCTSOF 1"1-INSULIN DEGRA-DATION BY PLACENTAL EXTRACT

Standard 15,000 X G placental extract was diluted fivefold with pH 7.4phosphate buffer to yield the equivalent of 16.6 mg. wet weight placenta perml. final reaction mixture. Final insulin concentration was as indicated inillustration. Incubation: 60 minutes, 380 C. Reaction mixture directly ap-plied for paper electrophoresis at end of incubation. Electrophoresis wasperformed at 40 C. to minimize further activity of the dilute enzymepreparation. Radioautograph of electrophoretograms is depicted above.

MIT and DIT markers (Figure 3). The ratios of"MIT-like"

45"DIT-like" radioactivity ranged from 3.0 to 4.5with four preparations of I131-insulin. Negligibleradioactivity was observed in the zones corre-sponding to ionic iodide. Further characteriza-tion of the radioactive products was not attemptedsince it is not germane to the present studies.Thus, whether the "MIT- and DIT-like" compo-nents consist entirely of iodinated tyrosines (anddo not include peptides of similar mobilities), andwhether all of the "origin" radioactivity representsunaltered I131-insulin cannot be said with certainty.However, from the data, simple deiodination maybe excluded as the basis for the manifest placentalactivity. Since the absolute formation of radio-active "MIT- and DIT-like" products was re-duced by the inclusion of increasing quantities ofunlabeled insulin in the reaction mixtures (Figure3), it was inferred that the 1131-insulin validly mir-rored the proteolysis of unlabeled insulin.

C. Liberation of TCA-soluble products of op-

tical density at 280 m1A. Documentation of pro-teolysis of unlabeled insulin was secured by themethod of Tomizawa and Williams (25). Ho-mogenates were prepared by the standard experi-mental procedure and dialyzed for 18 hours at 4°C. against repeated changes of 0.1 M phosphatebuffer, pH 7.4. Thereafter, 1 ml. aliquots wereincubated with 1 ml. of a) phosphate buffer or b)phosphate buffer containing 2 mg. crystalline in-sulin per ml. Control vessels containing phos-phate buffer alone or phosphate buffer plus insulinwere similarly incubated. At various times duringincubation, the contents of the beakers were quan-titatively decanted into 3 ml. of chilled 10 percent TCA, and the TCA-soluble supernatant solu-tions were read in a Beckman DU spectropho-tometer at 280 mu. Significant evolution of TCA-soluble, ultraviolet-absorbing moieties was onlyobserved in systems containing both insulin andplacental extract (Table IV).

The preliminary dialysis in these studies wasinstituted to reduce the appreciable concentration

121

NORBERTFREINKEL AND CHARLESJ. GOODNER

So/vent: BUTANOL- ACETIC

CotGpZrdsI RIGIN i i--I I-] . < I - ! , § t

_ _ L i INSULIN LOSug OmL. Without ExUcl. _. A, _ _ \,. _ z _ Jo -he ._ _ _, S .__ ._ _ w 1__ _. X ....._ . _ _ _ _ a ; _ ..................................... . . _ _ . . .__ ' 0 L < #--t -- - - -- - --- - -_ _ r BUTANOL- DIOXANE-AMMONIA

EZ1 Goair E]Hi 3

AL~~;~-4~~> INSULIN: 1.0#j/mL. + Extract

INSULIN: 41.0g./mL. + Extract --

FIG. 3. CHROMATOGRAPHYOF PRODUCTSOF I"'-INSULIN DEGRADATIONBY PLACENTAL EXTRACT

Reaction mixture: 2 mnl. pH 7.4 phosphate buffer containing 15,000 X G placental extract as

in standard experilictal procedure. Incubation: 30 minutes, 38° C. At end of incubation,aliquot of reaction mixture was applied to Whatman No. 1 filter paper, supplemented with car-

rier compounds, dried and chromatographed as indicated in text. Radioactive scans of chro-matograms are illustrated above.

of nonprecipitable components of optical densityat 280 mpt in the placental extracts. However,such dialysis also diminishes the insulin-degradingcapacity of placental preparations.

TABLE IV

Liberation of TCA-soluble products of opticaldensity at 280 mru.

Reaction mixture: 2 ml. pH 7.4 phosphate buffer :i crystalline, un-

labeled insulin in final concentration of 1 mg. per ml. + dialyzed15,000 X G placental extract as in standard experimental procedure.Incubation: 380 C.

Optical density at 280 my*

0 min. 30 min. 60 min.incuba- incuba- incuba-

Phosphate buffer plus tion tion tion

Phosphate buffer 0.075 0.070 0.072Insulin 0.080 0.075 0.078Placental extract 0.330 0.334 0.350Insulin plus placental extract 0.328 0.410 0.535

* Reaction mixture was decanted into TCA; opticaldensity of TCA supernatant was determined.

D. Bioassay. Alteration in the biological po-

tency of insulin incubated with placental extractswas assessed by injecting portions of control andreaction mixtures into the ear veins of fasted2 to 3 Kg. rabbits. Timed blood specimens forglucose analysis were obtained from the vesselsof the ear or by cardiac puncture. Each solu-tion was assayed in a separate animal. One offour experiments is presented in Table V. In-sulin which had been incubated for 30 minuteswith placental extract did not produce the same

degree of hypoglycemia as insulin which had beenincubated for a comparable period of time butwithout placental extract. The placental extracts

per se effected negligible changes in blood sugar

levels.E. Conclusions. The electrophoretic, chromato-

graphic and spectrophotometric studies demon-

I-

I:

122

PLACENTALINSULIN DEGRADATION

TABLE V

Inactivation of insulin by placental extracts

Reaction mixture: 2.0 ml. pH 7.4 phosphate buffer containing 15,000X G placental extract as in standard experimental procedure and

60 jig. (1.5 units) insulin per ml. Control vessels containing a)placental extract without insulin, and b) insulin without placentalextract, were incubated concurrently. Incubation: 30 minutes,380 C.

Assay: Rabbits were injected with 1.0 ml. of reaction mixture at end ofincubation.

Minutes after injectionReaction mixture (% of initial blood sugar)

30 min. 60 min.Phosphate buffer plus insulin 42.4 56.8Phosphate buffer plus placental 95.1 98.9

extractPlacental extract plus insulin 62.1 80.1

strate that the formation of TCA-soluble radio-activity from I13l-insulin by placental extractsrepresents proteolysis of the insulin molecule.The biological consequence of this proteolysis isevidenced by a reduction in the hypoglycemic po-tency of the insulin.

Non-Precipitoble( %of Total 1s3 )

25 .

20F

15

t0

5

INCUBATION (Minutes)

EXTRACT /5,000 g. Solubl SAmoteINSUL/N.' PO.02,mg. /mL.

U.)

I0 40

k 30

° 20

t 10

0.2 Q4 0.6 Q8 Q2 0.4PLACENTALEXTRACT (mL.)

FIG. 4. RELATIONSHIP BETWEEN QUANTITY OF PLA-CENTAL EXTRACT AND INSULIN DEGRADATION

15,000 X G placental extract was prepared as instandard experimental procedure. Various quantities ofunheated and boiled extract were incubated for 5, 10 or15 minutes in a reaction mixture consisting of 2.0 ml.volume and containing 200.02 ug. insulin per ml.

013

Vmw- 0.77 x IC- aK~ DEGRADATION-11.0,

SLOPE= Kmt11 ~ ~~~~Vmax.

Km = 3.8 x0oII

ldxli INSULIN M.W.:-A I I I I

)O7 mou/mmn.

),,g./mmn./gm.4.9 m/n. /Liter

r12 mo0.00i

12,000

1 2 3 4 5x 107 ( Liters/mol.)

S

FIG. 5. KINETICS OF INSULIN DEGRADATIONBY PLACENTAL EXTRACTS

Reaction mixture: 2 ml. pH 7.4 phosphate buffer containing various quantities of insulin and 15,000 X G placentalextract was prepared as in standard experimental procedure. Aliquots of reaction mixtures were removed at variousintervals during incubation. Nitrogen content of the 1: 3 15,000 X G placental extract employed for this study was

408.2 mg. per 100 ml.

123

EXTRACT: 15,000 x g. Soluble Supernate pH : 7.4 TEMP. : 38°C

REACTION MIXTURE: Extract from 0.083 gm. Placenta /mL.

I

- x 10INSULIN CONC. V

#9./VmL. ( minutes/mol. ).0.25 30p

NORBERTFREINKEL AND CHARLESJ. GOODNER

4. Characterization of the insulin-degrading sys-tems of human placenta

Although the 15,000 X G soluble supernatantfraction derived from human placentas does notrepresent a purified preparation, its activity wasassessed by various conventional techniques ofenzymology.

A. Relationship between insulin degradationand quantity of placental extract. Constantamounts of insulin were incubated with varyingamounts of the 15,000 x G placental extract. Arepresentative experiment is illustrated in Figure4. The formation of nonprecipitable I131 duringincubation of I131-insulin was directly correlatedwith the quantity of extract in the reaction mix-ture. Such a relationship did not obtain with ex-tract which had been boiled for 5 minutes prior toincubation.

B. Kinetics. The left-hand portion of Fig-ure 5 represents the fraction of J131-insulin de-graded as a function of time in the standard ex-perimental procedure. Prolonged linear degrada-tion was not observed at any substrate concentra-tion. Thus, for kinetic analysis, the initial velocity

20

I'.

I-

1o

5

0o40 50 6.7. 0 &0 9D 30 40 5 66D 70 8D 90

FINAL pH OF INCUBATION MIXTURE

FIG. 6. THE EFFECT OF PH ON INSULIN DEGRADATIONBY PLACENTALEXTRACT

Reaction mixture containing 2 ml. volume and 15,000 XG placental extract as in standard experimental procedure.Acetate buffer employed for pH 3.5 to 4.5; phosphatebuffer for pH 5.4 to 8.1; and Tris (hydroxymethyl)aminomethane buffer for pH 8.2 to 9.0 The pH values ofthe reaction mixtures before and after 15 minutes of incuba-tion were: 3.50, 3.97; 4.00, 4.40; 4.57, 4.87; 5.41, 5.91;6.00, 6.23; 6.65, 6.84; 7.00, 7.15; 7.40, 7.42; 8.10, 7.91;8.20, 8.27; and 9.00, 8.71.

TABL

Thermal inactivation of the iplacental

Reaction mixture: aliquots of 15,000at various temperatures (380,intervals (5, 10 and 20 minutesincubation was performed at 31buffer containing 100.1 jg. insextract as in standard experimec

Insulin degradation was asvessels after 1, 5, 15 and 30 minTCA-soluble radioactivity inextract.

Preincubationheating of Nonpre

extract during

Dura-Temp. tion

1min.

0 C. min.38 5 7.2

10 7.420 6.8

50 5 6.110 5.620 4.4

60 5 7.810 3.320 3.2

80 5 0.71020 0.5

,E VI at each concentration was determined as the prod-insulin-degrading capacity of uct of the initial insulin concentration and theI extracts zero time slope. A Lineweaver-Burk plot of the

values obtained during the first minute at the low-IX Gplacental extract were heated est concentrations of insulin is presented in the500, 600 and 801 C.) for varying

s) prior to incubation. Thereafter, right-hand portion of Figure 5. Durin this in-80 C. with 2 ml. pH 7.4 phosphate - o fFgr . uigti n3ulin per ml. and treated placental terval, degradation of insulin by boiled placentalntal procedure.sessed by sampling from reaction extracts is negligible and correction for the thermo-lutes of incubation and corrected forcontrol vessels incubated without stable activity does not affect the values signifi-

cipitable I131 (%) formed cantly. On the basis of a molecular weight ofincubation of extract for: 12,000 for insulin, the Michaelis-Menten constant

5 1s 30 (Km) for the thermo-labile reaction is 3.8 x 10-7mm. mln min.min. min. min. Moles per L. and Vmax is 0.77 x 10-7 Moles per

minute. Correction for the dilution in the stand-1426 15.8 21.8 ard experimental procedure yields a Vms1 of 11 jug.11.2 15.0 22.2 per minute per Gm. of wet weight of placenta.11.3 14.5 21.6 C. Stability. Placental extracts were heated11.1 13.2 22.1 for various intervals at 380, 500, 600 and 800 C.

8.7 10.8 18.2 and assayed for residual activity. As shown on7.8 9.1 14.4 Table VI, about one-fourth of the insulin-degrad-3.9 7.2 8.9 ing capacity which was observed during 15 to 30

2.9 3.5 6.7 minutes of incubation remained, even after 203.4 7.0 minutes of heating at 800 C.

1.6 3.6 6.2 The thermo-labile enzymatic activity was re-

INSULIN: 500.06 .G./mL.After

5 Min Incubation o-o15 * Incubation *-.

0 *\;

*-'1. . . . 0.

124

15~

PLACENTAL INSULIN DEGRADATION

duced by storing 15,000 x G placental extracts at- 180 C. for prolonged periods. Kinetic analysisindicated that this resulted from a diminution ofVmax without a significant change in Km. Thethermo-stable activity was not reduced by storageof frozen placental extracts.

D. Effect of pH and temperature. The rela-tionship between pH and enzymatic activity isshown in Figure 6. Experiments were conductedwith short incubation to minimize the contribu-tions of the relatively slow, thermo-stable compo-nent. At all insulin concentrations, insulin degra-dation was maximal between pH 6.5 and 7.0.

The relationship between insulin degradationand temperature of incubation is demonstrated inthe experiment presented in Table VII. During5 minutes of incubation, the mean temperaturevelocity constant was 22,000 cal. per Mole for therange 240 to 340 C. and 9,000 cal. per Mole for340 to 380 C. Reaction velocity declined at tem-peratures between 420 and 550 C.

E. Inhibition of enzymatic activity. Enzymeinhibitors previously assessed in the "insulinase"system of rat liver (10, 12, 26) were examinedwith placental extracts (Table VIII). Profoundinhibition was effected with reagents which mayaffect sulfhydryl groups by oxidation-reduction(cf., cupric salts), alkylation (cf., iodoacetate),or mercaptide formation (cf., p-chloromercuri-benzoate).

Specificity of the placental system was assessedby incubating small quantities of insulin (5 x 10-7

TABLE VII

Insulin degradation by placental extracts atvarious temperatures of incubation

Reaction mixture: 2.0 ml. pH 7.4 phosphate buffer containing unheatedor boiled 15,000 X Gplacental extracts as in standard experimentalprocedure and 25.01 jug. insulin per ml. Reaction was terminatedafter 5 minutes of incubation. Values for nonprecipitable 1131were corrected for control vessels incubated without placentalextract at identical temperatures.

Nonprecipitable 1131(% of total 11")

Unheated BoiledIncubation temp. extract extract

0 C.24.5 4.1 030.0 8.4 0.133.8 12.9 0.338.0 16.0 0.342.0 16.9 0.155.2 9.9 1.279.0 0.1 0.1

TABLE VIII

The effect of enzyme inhibitors on insulin degradation byplacental extracts

Reaction mixture: 2 ml. pH 7.4 phosphate buffer containing 100.03 jfg.insulin per ml., 15,000 X G placental extract as in standard experi-mental procedure A enzyme inhibitor. Incubation: 30 minutes,380 C.

Nonprecipi-Concentration table 1131

Inhibitor (X 10-3 M) (% of total 1131)

None 31.8p-Chloromercuribenzoate 0.5 2.9Cupric sulfate 0.5 5.4Zinc sulfate 0.5 25.1Jodoacetic acid 1.0 18.8Sodium azide 1.0 30.3Sodium arsenite 1.0 27.3

M)3 in the presence of 1 x 10-6 to 1 x 10-s MIconcentrations of crystalline bovine serum albu-min,' or protein hormones such as glucagon,l cor-ticotropin (ACTH),' thyrotropin (TSH),' soma-totropin (STH),' and lactogenic hormone(LTH) .1 Reactions were terminated after 5minutes of incubation to minimize the inhibitorypotential of products of proteolysis. At concen-trations of 1 x 10-6 M, none of the proteins ap-preciably altered the degradation of insulin (Fig-ure 7). Whenpresent in 20-fold excess, the pro-tein hormones effected a modest reduction in therate of insulin breakdown. This was most pro-nounced with ACTH (Figure 7). Unlabeled in-sulin, added to a final insulin concentration of1 x 10-6 to 1 x 10-5 M, more effectively reducedthe formation of nonprecipitable I131 than anyother protein (Figure 7).

Despite the relatively small effects of the pro-tein hormones, the potential physiological sig-nificance of the interrelationships prompted fur-ther examination of the inhibition. Results ofsuch studies with glucagon and ACTHare shownin Figure 8. During 30 minutes of incubation atmultiple concentrations, the effects of ACTHandglucagon conformed to a competitive type ofinhibition.

F. Conclusions. The rapid degradation of in-sulin by unboiled placental extracts exhibits thelability, kinetic characteristics, thermal relation-

3 Calculations of molar concentrations were based onthe following values for molecular weight: insulin, 12,000;bovine serum albumin, 70,000; glucagon, 4,200; cortico-tropin, 4,000; thyrotropin, 10,000; somatotropin, 44,250:and lactogenic hormone, 32,000.

125

NORBERTFREINKEL AND CHARLESJ. GOODNER

PROTEIN

CONCENTRATION

STH

TSH I X 106M

LTH

BOVINE ~.LB. .... .............BOVINE ALB.

0 5 10 15 20 25 30 35 40 45

NOWNPRECIPITABLEI131 (% OF TOTAL_ ,)

FIG. 7. THE EFFECT OF VARIOUS PROTEINS ON INSULIN

DEGRADATIONBY PLACENTAL EXTRACTS

Incubation: 5 minutes, 380 C. For details see text.

ships and circumscribed pH dependencies that are

typical of enzymatic reactions. Moreover, theactivity is impaired by sulfhydryl reagents andmay be competitively inhibited in the presence oflarge quantities of such protein hormones as glu-cagon or ACTH. Contrariwise, the degradationof insulin by boiled placental extracts does notappear to be enzymatically mediated. The activityis slow, stable, and relatively uninfluenced by vari-ation in the quantity of placental extract or by thetemperature of incubation.

5. Normiial valites for insulin degradation by hu-man placentas

In order to establish values for the normal in-sulin-degrading capacity of human placenta and toobtain control data for future studies of abnormalpregnancies, nine term placentas were assayed on

separate days with multiple concentrations of in-sulin and the standard experimental procedure.

Analyses were obtained following 5 and 30 min-utes of incubation. Five different preparations ofI13'-insulin were employed in the nine studies.Values for fractional insulin degradation in sys-

tems containing 25, 50, 100 and 400 /Ag. of crystal-line insulin per ml. are presented in Figure 9.Significant differences could not be demonstratedin the insulin-degrading capacities of placentas ob-tained by vaginal delivery and placentas obtainedat elective Caesarian section prior to the inception

of labor. Nitrogen content of the 15,000 x G su-pernatant solutions derived from these 1: 3 pla-cental homogenates averaged 423.9 mg. N per 100mln . ± 73.1.4 For a single experiment, a placentawas obtained at the spontaneous abortion of an18 week old fetus. An extract prepared from thisplacenta by the standard experimental procedurecontained 233.2 mg. N per 100 ml. and effecteddistinct degradation of insulin (Figure 9).

For purposes of comparison, four similar ex-periments were performed with livers from fedmale rats. Mean values for insulin degradation bythe 15,000 x G supernatant solutions of 1: 3 ratliver homogenates are shown in Figure 9. Theinsulin-degrading capacities of the rat liver prepa-rations exceeded those of comparably preparedplacental extracts. However, since the nitrogencontent of the four rat liver extracts averaged828.9 mg. per 100 ml., the differences in activitiesare not as great when expressed on a nitrogenrather than on a wet weight basis.

Conclusions. The studies with the nine termplacentas indicate that ranges of normal activity

EXTRACT. 15,000 x g Soluble Supe'noteINCUBA T0N. 30 Minutes

26 / Glucogon

22r / J l.O~~~~~~.0mg./mI.22 -ACTHJ

V 18

14 ,Glucogon1

10~~/ // z ~ACTH}

6-l - _ - - - ntrol

2~~~~~~~

SFIG. 8. THE NATURE OF THE INHIBITION OF PLA-

CENTAL INSULIN DEGRADATION IN THE PRESENCE OF

ACTHOR GLUCAGONReaction mixture: 2.0 ml. pH 7.4 phosphate buffer

containing 20.2, 40.2 or 100.2 Ag. insulin per ml., 15,000 X

G placental extract as in standard experimental proce-dure, and + 0.5 or 1.0 mg. ACTH or glucagon per ml.V = units of I"1'-insulin degraded in 30 minutes of incuba-tion at 38° C. as judged by the evolution of nonprecipita-ble I1`. S = units of insulin per ml. of reaction mixture.

4 Mean + standard deviation.

126

PLACENTALINSULIN DEGRADATION

INCGUSArOIN 5 MINUTES, 38°C 30 MINUTES, 38°CINSULIN 2 5 50 100 400 25 50 | 00 400

ftshcent/0 * v LCIECEAEA ETO;3-0Wt

90

70-

60 -

50-_0*0

i40 -a~r{M 5o/cpo, v,4, -

0~~~~~~

(0a 09 0 VAINL EIVRY TR

FIG. 9. THE DEGRADATIONOF INSULIN BY EXTRACTSOF HUMANPLACENTAAND LIVERS OF FED MALE RATS

Details of incubation and preparation of extracts are as in standard ex-perimental procedure. Values for placentas were derived from individualexperiments. Cross-hatched bars refer to mean values of four experimentswith 15,000 X G rat liver extracts.

are narrow and that the measurement of insulin-degrading capacity represents a reproducible pa-rameter of human placental function. The similarpotencies of both Caesarian section and "va-ginal delivery" placentas would suggest that in-sulin degradation does not result from an activa-tion of proteolytic mechanisms during labor.Rather, the manifest activity of the single 18 weekplacenta may constitute evidence that placentalmechanisms for insulin degradation are present atleast as early as the second trimester.

DISCUSSION

The present studies have provided the first de-tailed in vitro characterization of an insulin-de-grading system derived from human tissue. Ithas been demonstrated that cellular and subcel-lular preparations of human placenta can degradeJ131-insulin into radioactive products which aresoluble in TCA, electrophoretically distinct fromionic iodide, and which approximate the chro-matographic mobilities of iodinated tyrosines.This evidence for the proteolysis of 1181-insulin hasbeen corroborated and extended with unlabeled

crystalline insulin. Thus, it has been shown thatTCA-soluble, ultraviolet-absorbing products ofproteolysis are evolved during incubation of nativeinsulin with human placental extracts. Con-comitantly, there is an attenuation of the hypo-glycemic potency of the insulin.

Inactivation of insulin by extracts of certaintissues of animal origin has been known for sometime (26-29). With the advent of I113-insulin,the mechanisms have been investigated and char-acterized perhaps most extensively in rat and beefliver preparations (9-12, 20, 25, 30, 31). Thesimilarities between the published findings for ani-mal liver "insulinase" and the insulin-degradingsystem in human placenta are striking. Liver"insulinase," like the placental activity, a) is lo-calized in the soluble cytoplasm during subcellularfractionation (20, 29, 31, 32); b) is inhibited bysulfhydryl reagents (10, 12, 26); c) is reduced bydialysis or storage (28, 29, 33); and d) containsboth heat-stable as well as heat-labile components(10, 20, 26, 28, 32). Moreover, in animal liveras in human placenta, the heat-labile componentdisplays the activation energy, reaction kinetics

127

NORBERTFREINKEL AND CHARLESJ. GOODNER

and circumscribed pH dependencies which arecharacteristic of enzymatic systems (10, 12, 25, 29,30, 34). Indeed, the Michaelis-Menten constants,based on initial velocities of insulin degradationby rat liver (34) and human placental extracts,are of similar orders of magnitude; only the pHoptima differ. Although all the present studieswere conducted at an initial pH of 7.4 to simulatephysiological conditions, maximum placental ac-tivity is obtained at pH 6.5 to 7.0. Contrariwise,rat liver "insulinase" is most effective at pH 7.6to 8.0 (10). With the crude extracts of liver andplacenta which have been employed, these dif-ferences are difficult to interpret. However, itshould be mentioned that final pH values werenot recorded in any of the previous studies, andhence, the possibility of declining pH values dur-ing incubation of rat liver preparations cannot beexcluded.

The available data do not justify conclusionsregarding the substrate specificity of the heat-labile placental system, nor can it be statedwhether one or several enzymes participate in themanifest cleavage of insulin. Relative specificityis indicated by the fact that equimolar concentra-tions of a variety of proteins did not appreciablymodify reaction kinetics.5 For the moment, de-finitive dissection of the insulin-degrading systemof human placenta must await the results of fur-ther purification.

Regardless of its biochemical specificity, thephysiological implications of this system warrantconsideration. The placenta must be regarded, atleast, as yet another tissue which can both respondto insulin and degrade insulin. For the placenta,as for other such tissues, it remains to be dem-onstrated whether local hormonal degradation islinked to insulin action or whether degradation

5 The competitive inhibition observed with large quan-tities of ACTH and glucagon need not indicate a truesubstrate competition for the insulin-degrading system.In 18,000 X G extracts of rat liver, ACTHand glucagonalso retard insulin breakdown (25). However, thehepatic enzymes for degrading ACTHand glucagon maybe different from "insulinase" (35, 36) and "insulinase"in 18,000 X G rat liver extracts may be competitively in-hibited by products of the proteolysis of ACTHor glu-cagon (30). Thus, if separate enzymes for ACTHandglucagon proteolysis are also present in the 15,000 X Gplacental extracts, then similar interrelationships mightoccur.

merely conditions local hormonal activity byregulating the availability of unaltered insulin.In either event, the presence of the placenta as anadded site for the irreversible removal of insulinduring pregnancy could significantly alter the totalinsulin demands of the pregnant female. Unlessthere were commensurate reduction in the ex-trauterine insulin requirements, the maintenance ofhormonal equilibrium in pregnancy would necessi-tate the absolute availability of increased quanti-ties of insulin. At any given time in pregnancy,the magnitude of the placental contribution to ma-ternal insulin economy6 would be conditioned bya) the delivery of insulin to the placenta (i.e., pla-cental blood flow), and b) the potential insulin-degrading capacity of the placenta. The presentstudies have not documented the level of placentalactivity throughout human gestation, althoughmanifest degradation was observed with the single18 week placenta that was available for examina-tion. Concurrent experiments with rats have dem-onstrated that proteolysis of insulin can be effectedby the earliest formed rat placenta and that theactivity per milligram of nitrogen in placental ex-tracts remains constant throughout pregnancy(37). Thus, if rat data may be extrapolated tohumans, it may be inferred that the contributionsof placental insulin degradation to maternal insu-lin requirements should parallel a) the vascularityof the placenta, and b) the total functioning massof placental tissue. With this in mind, review ofthe clinical features of the diabetogenic effects ofpregnancy may be appropriate.

In human pregnancy, tolerance to intravenousglucose is not significantly altered (38) althoughthe response to exogenous insulin is diminished,especially in the latter half of gestation (33).Presumably, therefore, an increased endogenoussecretion of insulin is required for the preservationof normal carbohydrate metabolism during preg-nancy. This premise is supported by the clinicalsequence in subjects with diminished pancreaticreserve. Thus, frank diabetes mellitus frequentlysupervenes in prediabetic females during the latterphases of pregnancy, and, despite some conflicting

6 Since there are at present no experimental data rele-vant to the fetal elaboration or the transplacental passageof insulin in primate pregnancy, these factors have beenomitted from the above theoretical considerations.

128

PLACENTAL INSULIN DEGRADATION

reports (39), the maintenance requirements forinsulin of pregnant diabetics are often increasedat this time (4-8). Indeed, both Pedersen (5)and Bergqvist (7) observed the greatest aug-mentation of maintenance insulin requirementsduring the seventh month, a time at which theplacenta is nearly fully developed, senescent vas-cular changes in the placenta have not yet occurred,the heightened cardiac output of pregnancy is atpeak level (40), and, by inference, placental per-fusion is maximal. The expulsion of the con-ceptus is followed by a prompt reversion of thediabetogenic effects of pregnancy. Insulin re-quirement of diabetic patients may decline pre-cipitously in the immediate postpartum period andnormoglycemia may return in prediabetic sub-jects (4-8). Although the latter phenomenoncan be attributed in part to diminished food intakeand the heightened muscular effort of labor, stud-ies with intact rats in this laboratory would indi-cate that a true alteration in insulin metabolismoccurs (37). Near term, in the rat, there is anacceleration in the fractional rate of turnover oflabeled insulin which cannot be explained on thebasis of transplacental loss of intact insulin, andwhich disappears within hours following parturi-tion. Thus, from all the clinical evidence, somecorrelation does seem to obtain between insulinrequirements and the presence and the level offunction of an intrauterine insulin-degrading site.

However, alternative mechanisms could also ex-plain these clinical findings. Among the extra-uterine possibilities that have been suggested, thecase for contra-insulin factors of pituitary originis somewhat weakened by the recent studies of awoman who was hypophysectomized during preg-nancy (41). In this patient, while on constantthyroid and cortisone replacement therapy, glu-cose tolerance was abnormal and response to in-travenous insulin was diminished at 31 weeks ofgestation whereas both tests were normal 5 dayspostpartum. Similarly, the case for adrenal con-tra-insulin factors is hard to reconcile with the factthat levels of blood glucocorticoids remain elevatedin the immediate postpartum period (42) despitethe occasional acute reduction in insulin require-ments of diabetics. As a matter of fact, the wholeconcept of increased adrenocortical activity duringpregnancy has recently been challenged (43).

From all considerations, it would appear thatthe temporal relationships are most consonant withan intrauterine basis for the diabetogenic effectsof pregnancy. Mediation could be effected by a)placental (or even fetal) elaboration of contra-insulin factors or b) placental (or even fetal)degradation of insulin. The present studies do notaid in differentiating between the secretory and thedegradatory possibilities. Insofar as mechanismshave now been demonstrated for the proteolysis ofinsulin by placenta, an experimental basis has beenprovided for the latter possibility.7 However, suchdegradation need not be confined to insulin. Byanalogy to other tissues, the placenta might alsocontain enzymatic mechanisms for the proteolysisof glucagon or pituitary principles which affect pan-creatic cell activity. Moreover, the alternatives ofintrauterine secretion or degradation need not bemutually exclusive. It is doubtful whether anymetabolic change in as complex an event as mam-malian pregnancy can ever be explained in termsof a single mechanism. From the present data, itmay only be suggested that alterations in hormoneeconomy secondary to hormone degradation withinthe conceptus warrant consideration in any analy-sis of gestational changes in carbohydrate me-tabolism.

SUMMARY

1. The presence of a system for the proteolyticinactivation of insulin has been demonstrated inhuman placental tissue.

2. During incubation of insulin with cellularand subcellular preparations of human placenta,insulin is degraded into trichloroacetic acid-soluble,ultraviolet-absorbing products of proteolysis.Concomitantly, the hypoglycemic potency of insu-lin is attenuated.

3. The system has been characterized with I131-insulin and cell-free placental extracts. It hasbeen demonstrated that placental insulin-degradingactivity is localized within the soluble, nonparticu-

7 The placental insulin-degrading system may also pro-vide a means of studying one of the suggested etiologies ofdiabetes mellitus. Mirsky has postulated that increaseddegradation of insulin could antedate manifest diabetesmellitus and constitute the insulinogenic challenge in theprediabetic state (29, 30). The placenta may be proposedas an easily accessible human tissue wherein this thesismay be experimentally evaluated in subjects with largebabies or other prediabetic stigmata.

129

NORBERTFREINKEL AND CHARLESJ. GOODNER

late cytoplasm and consists of heat-stable and heat-labile components. The heat-labile activity effectsmost of the rapid inactivation of insulin and ex-hibits properties that are characteristic of enzy-matic mechanisms. At 380 C., and pH 7.4, theMichaelis-Menten constant (Km) for the reactionwas found to be 3.8 x 10-7 Moles per L., and Vmaxwas 11.0 /Ag. per minute per Gm. of wet weight ofplacenta.

4. The activities of multiple normal term pla-centas were assayed at various concentrations ofinsulin in order to obtain control data for futurestudies of abnormal pregnancies. It was demon-strated that insulin-degrading capacity constitutesa reproducible parameter of human placental func-tion. Manifest activity was displayed by a singleplacenta obtained at the spontaneous abortion ofan 18 week old fetus.

5. The insulin-degrading system of human pla-centa has been compared to liver "insulinase"preparations of animal origin.

6. It has been suggested that proteolytic in-activation of protein hormones within the con-ceptus may be of physiological significance in thegestational changes in carbohydrate economy.

ACKNOWLEDGMENTS

The excellent technical assistance of Miss WilmaBarshaw and the secretarial services of Miss Susan Colbyare gratefully acknowledged. The authors are also in-debted to Dr. Fred M. Snell for his suggestions in thestudies of enzyme kinetics and to Dr. Benjamin Tenney,Chief of the Obstetrical Service of the Boston City Hos-pital, and the members of his 1956-1958 resident staffwhose constant cooperation made this study possible.

REFERENCES

1. Freinkel, N., and Goodner, C. J. The metabolismof insulin by human placental tissue (abstract).J. clin. Invest. 1958, 37, 895.

2. Hurwitz, D., and Jensen, D. Carbohydrate metabo-lism in normal pregnancy. New Engl. J. Med.1946, 234, 327.

3. White, P. Symposium on diabetes mellitus; preg-nancy complicating diabetes. Amer. J. Med. 1949,7, 609.

4. Katsch, G. Schwangerschaft als Belastungsprobe furDiabetikerinnen. Zbl. Gynik. 1950, 72, 1756.

5. Pedersen, J. Course of diabetes during pregnancy.Acta endocr. (Kbh.) 1952, 9, 342.

6. Hoet, J. P., and Lukens, F. D. W. Carbohydrate me-tabolism during pregnancy. Diabetes 1954, 3, 1.

7. Bergqvist, N. The influence of pregnancy on dia-betes. Acta endocr. (Kbh.) 1954, 15, 166.

8. Hagbard, L. Pregnancy and diabetes mellitus. Actaobstet. gynec. scand. 1956, 35, suppl. I.

9. Mirsky, I. A., Perisutti, G., and Dixon, F. J. De-struction of I3"-labeled insulin by liver slices.Proc. Soc. exp. Biol. (N. Y.) 1954, 86, 228.

10. Mirsky, I. A., Perisutti, G., and Dixon, F. J. Thedestruction of I"'3-labeled insulin by rat liver ex-tracts. J. biol. Chem. 1955, 214, 397.

11. Tomizawa, H. H., Nutley, M. L., Narahara, H. T.,and Williams, R. H. Inactivation of insulin byrat liver extracts. J. Amer. chem. Soc. 1954, 76,3357.

12. Tomizawa, H. H., Nutley, M. L., Narahara, H. T.,and Williams, R. H. The mode of inactivationof insulin by rat liver extracts. J. biol. Chem. 1955,214, 285.

13. Gleason, Geoffrey I. Abbott Radio-Pharmaceuticals,Oak Ridge division. Personal communication.

14. Berson, S. A., Yalow, R. S., Bauman, A., Roth-schild, M. A., and Newerly, K. Insulin-I.31 metab-olism in human subjects: Demonstration of insu-lin binding globulin in the circulation of insulintreated subjects. J. clin. Invest. 1956, 35, 170.

15. Martin, D. B., Renold, A. E., and Dagenais, Y. M.An assay for insulin-like activity using rat adiposetissue. Lancet 1958, 2, 76.

16. Freinkel, N., Dowling, J. T., and Ingbar, S. H. Theinteraction of thyroxine with plasma proteins: Lo-calization of thyroxine-binding protein in Cohnfractions of plasma. J. clin. Invest. 1955, 34, 1698.

17. Dowling, J. T., and Goodner, C. J. Electronic ana-log integrator for simultaneous linear and integralgraphic data recording. In preparation.

18. Nelson, N. A photometric adaptation of the Somogyimethod for the determination of glucose. J. biol.Chem. 1944, 153, 375.

19. Ferrebee, J. W., Johnson, B. B., Mithoefer, J. C., andGardella, J. W. Insulin and adrenocorticotropinlabeled with radio-iodine. Endocrinology 1951, 48,277.

20. Narahara, H. T., Tomizawa, H. H., Miller, R., andWilliams, R. H. Intracellular distribution of an in-sulin-inactivating system of liver. J. biol. Chem.1955, 217, 675.

21. Burrows, B. A., Peters, T., and Lowell, F. C.Physical binding of insulin by gamma globulinsof insulin-resistant subjects. J. clin. Invest. 1957,36, 393.

22. Newerly, K., and Berson, S. A. Lack of specificity ofinsulin-I13-binding by isolated rat diaphragm.Proc. Soc. exp. Biol. (N. Y.) 1957, 94, 751.

23. Goodner, C. J., Ingbar, S. H., and Freinkel, N. Theinhibition of insulin degradation by plasma frac-tions of nondiabetic sera (abstract 71). Programof the fortieth meeting of the Endocrine Society,San Francisco, Calif. 1958.

130

PLACENTALINSULIN DEGRADATION

24. Kallee, E. Uber 13.J-signiertes Insulin, I Mitteilung(Nachweis). Z. Naturforsch. 1952, 7B, 661.

25. Tomizawa, H. H., and Williams, R. H. Studies onthe specificity of an insulin-inactivating system ofthe liver. J. biol. Chem. 1955, 217, 685.

26. Mirsky, I. A., and Broh-Kahn, R. H. The inactiva-tion of insulin by tissue extracts. I. The distribu-tion and properties of insulin inactivating extracts.(Insulinase.) Arch. Biochem. 1949, 20, 1.

27. Schmidt, A. A., and Saatchian, R. L. Effect of sometissues of the animal organism on the activity ofinsulin. 2. exp. Biol. med. 1929, 11, 42, as quotedin Chem. Abst. 1930, 24, 4859.

28. Lehmann, H., and Schlossmann, H. The action ofcell-free muscle extract on insulin. J. Physiol.1938, 94, 15P.

29. Mirsky, I. A. The insulinase and insulinase-inhibitoractivity of the liver. Ciba Found. Coll. Endocr.1953, 6, 263.

30. Mirsky, I. A. Insulinase, insulinase-inhibitors, anddiabetes mellitus. Recent Progr. Hormone Res.1957, 13, 429.

31. Vaughan, M. The inactivation of insulin by an en-zyme from rat liver. Biochim. biophys. Acta 1954,15, 432.

32. Beyer, R. E. A study of insulin metabolism in aninsulin tolerant strain of mice. Acta endocr.(Kbh.) 1955, 19, 309.

33. Burt, R. L. Peripheral utilization of glucose inpregnancy. III. Insulin tolerance. Obstet. andGynec. 1956, 7, 658.

34. Yalow, R. S., and Berson, S. A. Apparent inhibitionof liver insulinase activity by serum and serum

fractions containing insulin-binding antibody. J.clin. Invest 1957, 36, 648.

35. Mirsky, I. A., and Perisutti, G. The relative speci-ficity of the insulinase activity of rat liver ex-tracts. J. biol. Chem. 1957, 228, 77.

36. Tomizawa, H. H., and Halsey, Y. D. Isolation ofan insulin-degrading enzyme from beef liver. J.biol. Chem. 1959, 234, 307.

37. Goodner, C. J., and Freinkel, N. The effect of preg-nancy on insulin metabolism in the rat. Clin. Res.1959, 7, 247.

38. Burt, R. L. Peripheral utilization of glucose inpregnancy and the puerperium. Obstet. andGynec. 1954, 4, 58.

39. Patterson, M., and Burnstein, N. Diabetes and preg-nancy. A clinical analysis. Arch. intern. Med.1949, 83, 390.

40. Bader, R. A., Bader, M. E., Rose, D. J., and Braun-wald, E. Hemodynamics at rest and during exer-cise in normal pregnancy as studied by cardiaccatheterization. J. clin. Invest. 1955, 34, 1524.

41. Little, B., Smith, 0. W., Jessiman, A. G., Selenkow,H. A., Van't Hoff, W., Eglin, J. M., and Moore,F. D. Hypophysectomy during pregnancy in apatient with cancer of the breast: Case reportwith hormone studies. J. clin. Endocr. 1958, 18,425.

42. Bayliss, R. I. S., Browne, J. C., Round, B. P., andSteinbeck, A. W. Plasma-17-hydroxycorticoster-oids in pregnancy. Lancet 1955, I, 62.

43. Migeon, C. J., Bertrand, J., and Wall, P. E. Physio-logical disposition of 4-C14-cortisol during latepregnancy. J. clin. Invest. 1957, 36, 1350.

131