Degradation of Collagen by a Human

Granulocyte Collagenolytic System

GERALDS. LAZARUS, JOHNR. DANIELS, ROBERTS. BROWN,HowARDA. BLADEN, and HAROLDM. FuLLMERFrom the Laboratory of Histology and Pathology and the Laboratory ofBiochemistry of the National Institute of Dental Research, and theMedicine Branch of the National Cancer Institute,Bethesda, Maryland 20014

A B S T R A C T This report suggests a mechanismfor collagen degradation mediated by human gran-ulocytic leukocytes. A specific collagenase, whichis extractable from human granulocytes, has beenpartially purified by DEAE chromatography.This collagenolytic enzyme is operative at physio-logical pH and is inhibited by EDTA, cysteine,and reduced glutathione but not by human serum.The enzyme cleaves the collagen molecule into twospecific products, without loss of helical conforma-tion. Electron micrographs of segment long spac-ing aggregates indicate that the cleavage occursone-quarter of the length from the carboxy termi-nal end of the molecule. Experiments with crudeextracts from granulocytes suggest that the spe-cific products of granulocyte collagenase activityare then degraded by other proteases present inthe human granulocyte.

INTRODUCTION

Degradation of collagen fibrils and collagen-con-taining structures such as basement membrane oc-curs in inflammatory responses concomitant withleukocyte infiltration (1). The precise mechanismsinvolved have required clarification since nativecollagen is resistant to hydrolysis by many proteo-lytic enzymes. Recently we have detected a colla-

Dr. Lazarus' present address is Department of Derma-tology, Massachusetts General Hospital, Boston, Mass.Dr. Brown's present address is Yale New HavenMedical Center, New Haven, Conn.

Received for publication 27 June 1968 and in revisedform 15 August 1968.

genolytic enzyme which is directly extractablefrom the granule fraction of human granulocyticleukocytes (2). This enzyme cleaves the collagenmolecule into two distinctive products, which aresimilar to those produced by collagenases derivedfrom tadpole skin (3-7), human skin (8-10), andsynovium (11, 12), and the postpartum rat uterus(13). Unlike these and similar collagenolytic ac-tivities released from human gingiva (14-16) andbone (17, 18), granulocytes yield collagenase onextraction and do not require tissue culture to pro-duce detectable enzyme. The only other extractablecollagenolytic factor described has been an acidhydrolase found in rat bone (19). In this reportwe detail the -partial purification of the enzyme,describe some of its properties, and suggest amechanism by which human granulocytes mediatecollagen degradation.

METHODSEnzyme preparation. The white cells were obtained

from 500 ml units of fresh whole blood which weredrawn from normal adult donors. The units of bloodwere centrifuged at 1500 g for 3 min and the 60 ml whitecell-rich interface between the plasma and packed redblood cells was collected. Six of these white cell-rich unitswere combined and added to an equal volune of solu-tion containing 0.45% of sodium chloride and 1.5% ofdextran (mol wt 186,000). After allowing the red cellsto settle for approximately 1/2 hr at room temperaturethe white cell-rich supernatant was decanted and cen-trifuged at 1500 g for 8 min. All subsequent steps werecarried out at 4°C. The cell button was resuspended insaline and the remaining red cells removed by briefhypotonic hemolysis (20). The white cells were thenwashed three times with saline and collected by centri-

2622 The Journal of Clinical Investigation Volume 47 1968

fugation at 400 g for 10 min. This preparation wasevaluated by cell counts and differential smears. Thecells were then homogenized by hand in Tyrodes solutionusing a ground glass homogenizer. After the homogenatewas repeatedly frozen and thawed it was centrifuged at25,000 g for 15 min. The supernatant was dialyzed for 1hr against 0.002 M calcium chloride in 0.01 M Tris buffer,pH 8.5, and then clarified by centrifugation.

The crude extract was purified by chromatography ondiethylaminoethyl cellulose 1 in a water-jacketed column(1.7 cm diameter by 7 cm length) at I C. The elutingbuffer was 0.01 M Tris, pH 8.5, containing 0.002 M cal-cium chloride on which was superimposed a linear gradi-ent from 0.0 to 0.15 M sodium chloride over a total vol-ume of 400 ml. The flow rate was 50 ml/hr and 5-mlaliquots were collected. Optical density at 280 m,. wascontinuously monitored. Activity was located by acryla-mide gel electrophoresis of the products of incubationsof eluate with collagen at 250C for 24 hr.

Protein determinations were done using the Millermodification of the Lowry technique (21) and caseino-lytic activity was assayed using the method of Nagai,Lapiere, and Gross (5).

Substrate preparation. The acid-extracted radioactivecollagen was prepared after injecting 40-day-old SpragueDawley rats intraperitoneally with uniformly labeled 14C-glycine. Each rat received 50 /Ac of the isotope 72, 60, 48,24, and 12 hr before sacrifice. The acid-extractable col-lagen was purified as described by Kang, Nagai, Piez,and Gross (6). Resistance to nonspecific degradationwas evaluated by incubating it with a number of differentproteolytic enzymes in the 14C-labeled collagen fibril as-say. The collagen was found to be almost completely re-sistant to solubilization by enzymes other than clostridialcollagenase. Indeed, when collagen fibrils were incubatedwith trypsin (50% w/w) only 5% of the substrate wassolubilized.

Collagenase assays. Four methods of detection of col-lagenolytic activity were used: (a) release of radioactivedegradation products from reconstituted '4C-glycine-la-beled collagen fibrils (5), (b) viscometry at 25°C, (c)acrylamide gel electrophoresis (7), and (d) electron mi-croscopy of segment long spacing aggregates of collagen(4).

The reconstituted '4C-labeled collagen fibrils were pre-pared as follows. 2 mg of collagen were dissolved per mlof distilled water by gentle stirring overnight at 5°C.After the solution was dialyzed for 12 hr against 0.2 Msodium chloride in 0.05 M Tris buffer, pH 7.6, 5°C, it wascentrifuged at 65,000 g for 2 hr. 200-,ul aliquots of the su-pernatant (containing 1150 dpm) were pipetted into3-ml test tubes and allowed to gel at 37°C overnight. Tothe opalescent gel, which consisted of organized collagenfibrils, was added 0.25 ml of 0.001 M calcium chloride in0.05 M Tris buffer, pH 7.6, and 0.5 ml of the sample tobe assayed. The mixture was incubated for 18 hr at 37°Cand then filtered through a 0.9 ,u pore size, 13 mmdiam-

1 Whatman DE 52, H. Reeve Angel and Co., Inc.,Clifton, N. J.

eter Versapor 2 filter in a Swinney adapter. The filter re-tained insoluble collagen fibrils. Solutions containing in-tact collagen molecules and peptide reaction productspassed through the filter and were counted. A 0.5 ml ali-quot of the filtered solution was added to 20 ml of Bray'ssolution (22) and counted in a liquid scintillation spec-trometer. When serum was included in the incubationmixtures the sample precipitated in Bray's solution. Thiswas circumvented by treatment of the sample with 0.5 mlof 1.0 M sodium hydroxide at 60°C for 2 hr before addi-tion of the counting mixture. Counts were determined towithin a 2%o error, and after correction for quenching bythe channels ratio method, they were adjusted to 100%efficiency (23). All experiments included assays of bufferblanks. Concomitant trypsin 3 controls (25 ug) indicatedthe possible extent of nonspecific proteolytic breakdown ofthe collagen gel.

Kinetic experiments which measured the decrease inviscosity of collagen solutions with cleavage of the mole-cule were performed at 25°C in Ostwald viscometers(flow times 70-95 sec for 1.5 ml of water at 25°C). The6 ml viscometry solution included 1 ml of enzyme, andcontained final concentrations as follows: collagen 0.67mg/ml, calcium chloride 0.015 M, sodium chloride 0.35 M,and Tris buffer 0.05 M, pH 8.5. The influence of pH wasstudied by using appropriate Tris-HC1 or Tris-maleatebuffers. All experiments were done in duplicate and in-cluded buffer blanks. Results were computed as the percent reduction in specific viscosity with time from the ini-tial measurement. Parallel optical rotation measurementswere made on several viscosity experiments by followingduplicates of the reaction mixture in a spectropolarimeter(model 80, 0. C. Rudolph & Sons, Inc., Caldwell, N. J.)with an oscillating polarizer at 313 m/u at 250C.

Both viscometry and the radiofibril assay were used tostudy potential modifiers of collagenase activity. Reducedglutathione,4 L-cysteine,5 and disodium EDTA were dis-solved in the experimental buffers to make a final con-centration of 0.01 M at pH 8.0 in the incubation mixtures.Similarly, in other reaction mixtures various dilutionsof normal pooled human serum were added in volumesequal to that of the enzyme preparation.

At the completion of all viscometry experiments theproducts were studied by acrylamide gel electrophoresis.Reaction mixtures were precipitated by increasing thesodium chloride concentration to 20o (wt/vol). The pel-let formed by centrifugation at 65,000 g for 15 min wasdissolved in 8 M urea titrated to pH 5.3 with HCL.After dialysis against the pH 5.3 urea, the solution wassubjected to electrophoresis according to the method ofSakai and Gross (7). Comparable results were obtainedwhen, in other experiments, salt precipitation wasomitted.

2 Gelman Instrument Co., Ann Arbor, Mich.3 Trypsin, 2 X crystallized, Worthington Biochemical

Corp., Freehold, N. J.4 Sigma Chemical Co., St. Louis, Mo.5 L-cysteine hydrochloride, Nutritional Biochemicals'

Corp., Cleveland, Ohio.

Granulocyte Collagenolytic System 2623

1.o0 lulose column is presented in Fig. 1. Collagenaseactivity, as measured by both radiofibril and acryl-

Absorbance,?80maamide gel electrophoresis assays, was confined to

Absorbonce 280m~a 60 ml eluted over a sodium chloride concentration08 _ range of 0.045-0.075 M. The fraction with maxi-

mumcollagenase activity was free of caseinolyticactivity and had a protein content of 0.6 mg/ml.The degree of purification could not be determined

20.6 _ | 1t1 u 60 since the crude preparation was not stable untilE other leukocyticproteases were removed byTDEAEN lnl l \ z chromatography. decrease in specific viscosity was effected by them | partially purified granulocyte collagenase when the

ipSRsodiofibri/ OSSOv, q < reaction went to completion. No change in opticalI I DPMreleasedt z rotation could be detected when viscosity and~~~~~~~~~

0.2 _-i 20 M polarimetry were observed in parallel experi-vLFay~Em/de Aecfivllyoresb ments (Fig. 2). This indicated that the over-allIocrylamide elec/rophoresis

helical structure of the products had been main-/ \ ~~~~~~~~tained.

0.0/2 300 The activity of the partially purified enzyme asEFFLUENT VOLUME,ml a function of pH was also studied by viscometryFIGURE 1 Purification of collagenolytic activity from (Fig. 3). Maximal activity was found close to pH8 X 108 granulocytes on a DEAE column at 1VC. Theeluting buffer was 0.002 M calcium chloride in 0.01 M O A A ATris, pH 8.5, on which was superimposed a linear saltgradient. Shown are optical density at 280 mg& and col-lagenolytic activity as measured by acrylamide gel elec-trophoresis and reconstituted radioactive collagen fibrilassays. 20

To study the morphology of the collagen moleculesafter granulocyte collagenase cleavage, segment longspacing collagen (SLS) was prepared from standard C 40viscometry mixtures according to the method of Gross <and Nagai (4). Electron micrographs were taken usinga Siemens Elmiskop I electron microscope with double Zcondenser illumination and a 50 A aperture at 40,000-80,000 "magnification.

0-

RESULTS

The granulocyte collagenase was extracted from 80a heterogeneous population of white cells withoutregard to lymphocyte contamination, because thesecells have been shown not to contain collagenase(2). 100

An extract of 1010 white cells (80% granulo- TM2 4 20cytes) yielded 350 mg of protein. 17 mg of thiscrude extract was able to solubilize 28%o of a ra- FIGURE 2 Parallel determinations of per cent changein optical rotation (A-A) and per cent decrease indioactive collagen fibril gel and had caseinolytic ac- specific viscosity (0-0) during incubation of collagentivity equivalent to 50 ,ug of trypsin. Chromatog- with the partially purified granulocyte collagenase withraphy of this material on a diethylaminoethyl cel- time (pH 8.5).

2624 Lazarus, Daniels, Brown, Bladen, and Fullmer

w

1L 60.4

- 50O) v0 NF' 40eOw

- 20

210a$)

F-_

6 7 8pH

FIGURE 3 Activity of the partially puricollagenase as a function of pH. Activit3the per cent decrease in specific viscosit250C.

7.8. Little activity could be detectedpH 6.5 or above pH 9.5.

The acrylamide gel electrophoremthe denatured products of incubati'with both the purified and crude entions are shown in Fig. 4. The cconsists of monomers (a), dimers (Imolecular weight species. Incubati(

with the partially purified granulocyte collagenaseresults in discrete products of lower molecularweight which are similar to those seen with otheranimal collagenases (6, 7, 10). The products aredenoted by superscripts A and B which are theN-terminal 3/4 and C-terminal 1/4 of the mole-cule respectively (6). The double bands for eachspecies reflect the heterogeneous chain structure ofthe collagen molecule (6). The complex patternproduced after -incubation with the crude extractis also presented for comparison. The numerousbands indicate extensive hydrolysis of the entire

9 -- - molecule.

ified granulocyte Fig. 5 is an electron micrograph of segment longy is measurec as spacing collagen (SLS) aggregates. On the righty after 20 hr at is SLS formed from collagen which had been in-

cubated with heat-inactivated purified granulocyted either below collagenase and shows the usual length (2700-

2800 A) and periodicity. On the left is SLS col-

sis patterns of lagen formed after incubation with active granu-on of collagen locyte collagenase. The carboxy terminal one-zyme prepara- quarter of the molecule has been cleaved off. Theontrol pattern shortened molecules (2150 A) were all of the same,8), and higher length and no SLS between 2150 and 2750 A wason of collagen seen. The electron micrographic site of cleavage

alaA

DegrodativeProducts

-BufferFront

CO:N O..

CON TROL

- _...,.... a_

.::.: :.... ....:

::::::: .: A::!: :..: ::! :: It .. a....::.

:.

....': : . : ...:.

* .... . A:.REACTIONMIXTURE

FIGURE 4 Acrylamide gel electrophore-sis patterns of products from incubationsof collagen with heat inactivated granylo-cyte collagenase (left), partially purifiedgranulocyte collagenase (middle), and un-purified granulocyte collagenase (right).a, monomeric chain; ,f, dimeric chain; su-perscript A, N-terminal 3/4 cleavageproduct; superscript B, c-terminal 1/4cleavage product.

REACTIONMIXTURE

Granulocyte Collagenolytic System 2625

appears identical with that of the tadpole col-lagenase (4).

Collagenase activity was inhibited by boiling theenzyme for 5 min, addition of sodium EDTAto afinal concentration of 0.01 M, omission of calcium,or addition of reduced glutathione or cysteine to afinal concentration of 0.01 M (Fig. 6). Becauseserum has been shown to inhibit collagenase ac-tivity derived from other tissues (8) its effect onthe granulocyte enzyme was studied. A clear dif-ference was observed: the addition of serum, orserum diluted 1: 10 with buffer, to the partiallypurified enzyme did not inhibit viscosity fall or

FIGURE 5 Comparison of SLS aggregates from colla-gen incubated with active partially purified granulocytecollagenase (left) and heat inactivated granulocyte col-lagenase (right). The SLS formed after incubation withactive enzyme is missing the carboxy terminal 1/4 of themolecule.

CONTROL ENZYME ENZYME+Cysteine, GSH,

EDTA, BoilingFIGURE 6 The effect of cysteine 0.01 M, reduced glu-tathione (GSH) 0.01 M, EDTA 0.01 M, and boiling onpurified granulocyte collagenase activity as determinedby acrylamide gel electrophoresis (pH 8.5, 250C, 20 hr).

effect the acrylamide gel electrophoresis pattern(Fig. 7). Furthermore, the addition of serum tothe crude enzyme preparation prevented produc-tion of multiple digestion products and resulted inan acrylamide gel electrophoresis pattern similarto that produced by the partially purified enzyme.A crude granulocyte preparation which effected aviscosity fall of 25%o in 4 hr was able to solubilize28% of a collagen gel while a sample of the puri-fied granulocyte enzyme which reduced viscosity55% in 4 hr solubilized only 17% of the collagenfibrils (Table I). When the effect of serum onsolubilization of radioactive collagen fibrils by thecrude and partially purified enzyme preparationswas studied, partial inhibition was found. The ap-parent lack of correlation between these two as-says will be discussed below.

DISCUSSION

The granulocyte enzyme cleaves the collagenmolecule into two specific pieces. These distinctiveproducts of collagenase action (/8A, aA, and aB)are shown in the acrylamide gel electrophoresispatterns of the denatured reaction mixtures.Electron micrographs demonstrate the largerproduct of the cleavage which is the N-terminal

2626 Lazarus, Daniels, Brown, Bladen, and Fullmer

d]s...]a

a3al

*.w ..

... -...........

N

CONTROL

aA [

aA [

S C

aB I

c:6 Id oBE e

pk>°j:: ':: Sm :°*.:

:: ,.

,2:, ..:v: :: isat

D P

PURIFIEDENZYME+ SERUM

CRUDEENZYME

AI /3A

hAI aA

CRUDEENZYME+ SERUM

FIGURE 7 Acrylamide gel electrophoresis patterns of the products from incubations of collagen with partially puri-fied and crude granulocyte collagenase with and without human serum (pH 8.5, 20 hr). a, monomeric chain; fi, di-meric chain; superscript A, N-terminal 3/4 cleavage product; superscript B, C-terminal 1/4 cleavage product; DP,nonspecific degradation products; SC, serum components.

3/4 of the molecule. Cleavage of the collagen mole-cule does not cause any major change in helicalstructure of the products since the initial negativeoptical rotation is maintained. The enzyme is ac-

TABLE IEffect of Normal Human Serum and Reduced Glutathione

(GSH) on Crude and Partially Purified GranulocyteCollagenase Activity as Measured by the Radio-

active Reconstituted Collagen Fibril Assay

Collagen gelDPM solubilized

Crude granulocyte collagenase 300 28 =1:2Crude granulocyte collagenase

+ serum 125 11 41Purified granulocyte collagenase 175 17 -4-1.5Purified granulocyte collagenase

+serum 90 941Purified granulocyte collagenase

+GSH0.O1M 35 4+1Trypsin 50 ,ug s0 5 ::I

Results presented have been corrected for the buffer blankof 75 DPM. Each gel initially contained 0.4 mg collagen(1150 DPM). Values expressed are the mean of four deter-minations +SD.

tive over a physiological pH range. The pH profilepresented is semiquantitative in that no attempthas been made to establish reaction rates at satu-rating substrate concentration. The enzymatic ac-tivity can be blocked by EDTAand the free sulf-hydryl compounds cysteine and reduced glutathi-one. These properties are shared with collagenasesfrom other tissue sources (3-8, 11-13).

There are, however, two major differences be-tween the granulocyte collagenase and other re-ported animal collagenolytic enzymes of this type.First, the enzyme is readily detected on extractionof the cells of origin and is apparently stored in theleukocyte granule (2). Collagenolytic activity fromother sources requires tissue culture for detection(3, 8-14). Its presence in the granulocyte is dis-tinctive since attempts to extract similar activityfrom comparable numbers of human lymphocytes(2) and rabbit alveolar macrophages 6 have failed.

Second, granulocyte collagenase is not inhibitedby human serum. When specific cleavage of thecollagen molecule is studied by viscometry andacrylamide gel electrophoresis, no inhibition of

6Lazarus, G. S., and J. Goggins. Unpublished data.

Granulocyte Collagenolytic System 2627

granulocyte enzyme action is found in the presenceof serum. This is at variance with observationsmade with other collagenases since all these enzymesare inhibited by serum when studied by identi-cal techniques (8, 24). In contrast, the addition ofserum to either the crude or partially purifiedgranulocyte enzyme inhibits dissolution of collagenfibrils when the radioactive fibril assay is em-ployed. Because of the observations on collagenin solution, this finding cannot be ascribed to in-terference with the specific cleavage of the mole-cule into A and B pieces. Furthermore, the partiallypurified granulocyte collagenase is less effectivein solubilizing collagen fibrils than comparableamounts of the crude granulocyte collagenaseas measured by viscometry. These observationssuggest that the specific cleavage of the mole-cule is inefficient in dissolving the fibril and sec-ondary proteolytic activity which is inhibitableby serum, facilitates effective solubilization. Thisimpression was strengthened by observations madewith the crude granulocyte extract. When incuba-tion mixtures using this extract were studied, notonly was solubilization of collagen fibrils marked,but numerous products were seen on acrylamidegel electrophoresis. Upon addition of serum, how-ever, only products specific for granulocyte col-lagenase were found: this condition was associatedwith a decrease in dissolution of radioactive col-lagen fibrils. Thus, these experiments demonstratethe presence of serum-inhibitable protease ac-tivity in the crude granulocyte preparation whichincreases the effectiveness of granulocyte collage-nase in solubilizing collagen fibrils. Whether thesertni-inhibitable, solubilizing activity of the par-tially purified granulocytic collagenase is a prop-erty of the enzyme itself or results from persistentcontamination by other enzymes cannot yet bestated.

Sakai and Gross (7) have shown that the prod-ucts of tadpole collagenase action are more sus-ceptible to tryptic hydrolysis than the intact mole-cule. Thev correlated this observation with thelowered melting points of the specific products.That the specific cleavage products of the granu-locytic enzyme can be further degraded by otherproteases in the white cell is shown by the nu-merous bands found on acrylamide gel electro-phoresis after action by the crude granulocytepreparation. These observations suggest that dur-

ing granulocyte-mediated dissolution of collagenfibrils the primary effect of the collagenase is tocleave the collagen molecule into two fragmentswhich are then more susceptible to hydrolysis byother proteases.

ACKNOWLEDGMENTWe wish to thank Dr. Karl Piez for his many helpfulsuggestions and Miss Jane Lian for her outstandingtechnical assistance.

REFERENCES

1. Cochrane, Ch. G. 1967. Mediators of the arthus andrelated reactions. Progr. Allergy. 11: 1-35.

2. Lazarus, G. S., R. S. Brown, J. R. Daniels, andH. M. Fullmer. 1968. Human granulocyte collagenase.Science. 159: 1483.

3. Gross, J., and C. M. Lapiere. 1962. Collagenolyticactivity in amphibian tissues: a tissue culture assay.Proc. Natl. Acad. Sci. U. S. 48: 1014.

4. Gross, J., and Y. Nagai. 1965. Specific degradation ofthe collagen molecule by tadpole collagenolytic en-zyme. Proc. Natl. Acad. Sci. U. S. 54: 1197.

5. Nagai, Y., C. M. Lapiere, and J. Gross. 1966. Tad-pole collagenase, preparation and purification. Bio-chemistry. 5: 3123.

6. Kang, A. H., Y. Nagai, K. A. Piez, and J. Gross.1966. Studies on the structure of collagen using acollagenolytic enzyme from tadpole. Biochemistry. 5:509.

7. Sakai, T., and J. Gross. 1967. Some properties of theproducts of reaction of tadpole collagenase with col-lagen. Biochemistry. 6: 518.

8. Eisen, A. Z., J. J. Jeff rey, and J. Gross. 1968. Humanskin collagenase isolation and mechanism of attacton the collagen molecule. Biochem. Bio phys. A cta.151: 637.

9. Fullmer, H. M., G. Lazarus, W. A. Gibson, A. C.Stam, Jr., and C. Link. 1966. Collagenolytic activity ofthe skin associated with neuromuscular diseases in-cluding amyotrophic lateral sclerosis. Lancet. 1: 1007.

10. Riley, W. B., Jr., and E. E. Peacock, Jr. 1967. Identi-fication, distribution and significance of a collageno-lytic enzyme in human tissue. Proc. Soc. Exptl. Biol.MIed. 124: 207.

11. Evanson, J. M., J. J. Jeffrey, and S. M. Krane. 1967.Human collagenase: Identification and characterizationof an enzyme from rheumatoid synovium in culture.Science. 158: 499.

12. Lazarus, G. S., H. M. Fullmer, C. H. Oliver, C. V.Multz, W. F. Barth, J. L. Decker, and E. J. Kamin.1967. Collagenolytic activity of human rheumatoidsynovium. Proceedings of the 4th Pan-American Con-gress of Rheumatology. Excerpta Medica Foundation,New York. 19.

13. Jeffrey, J. J., and J. Gross. 1967. Isolation and char-acterization of a mammalian collagenolytic enzyme.Federation Proc. 26 (2) : 670.

2628 Lazarus, Daniels, Brown, Bladen, and Fullmer

14. Fullmer, H. M., and W. Gibson. 1966. Collagenolyticactivity in gingivae of man. Nature. 209: 728.

15. Beutner, E. H., C. Triftshouser, and S. P. Hazen.1966. Collagenase activity of gingival tissue frompatients with peridontal disease. Proc. Soc. Exptl.Biol. Med. 121: 1082.

16. Bennick, A., and A. M. Hunt. 1967. Collagenolyticactivity in oral tissues. Arch. Oral Biol. 12: 1.

17. Walker, D. G., C. M. Lapiere and J. Gross. 1964.A collagenolytic factor in rat bone promoted by para-thyroid extract. Biochem. Biophys. Res. Comin. 15:397.

18. Fullmer, H. M., and G. Lazarus. 1967. Collagenasein human, goat and rat bone. Israel J. Med. Sci. 3:758.

19. Woods, J. F., and G. Nichols, Jr. 1965. Collagenolyticactivity in rat bone cells. J. Cell Biol. 26: 747.

20. Fallon, H. J., E. Frei III, J. D. Davidson, J. S. Trier,and D. Burk. 1962. Leukocyte preparations from hu-man blood: evaluation of their morphologic and meta-bolic state. J. Lab. Clin. Med. 59: 779.

21. Miller, G. L. 1959. Protein determination for largenumbers of samples. Anal. Chem. 31: 964.

22. Bray, G. A. 1960. A simple efficient liquid scintilla-tor for counting aqueous solutions in a liquid scintil-lation counter. Anal. Biochem. 1: 279.

23. Krichevsky, M. I., S. A. Zaveler, and J. Bulkeley.1968. Computer-aided single or dual isotope channelsratio quench correction in liquid scintillation counting.Anal. Biochem. 22: 442.

24. Lazarus, G. S., J. L. Decker, C. H. Oliver, C. V.Multz, J. R. Daniels and H. M. Fullmer. 1968. Col-lagenolytic activity of rheumatoid synovium. N. Engl.J. Med. In press.

Granulocyte Collagenolytic System 2629