Vol. 53 BIOLOGICAL ASSAY OF HYALURONIDASE 59

REFERENCES

Bacharach, A. L., Chance, M. R. A. & Middleton, T. R.(1940). Biochem. J. 34, 1464.

Duran-Reynals, F. (1946). Bact. Rev. 6, 197.Hechter, 0. (1947). J. exp. Med. 85, 77.Hechter, 0. (1950). Ann. N.Y. Acad. Sci. 52, 1028.

Humphrey, J. H. (1943). Biochem. J. 37, 177.McClean, D. (1942). J. Path. Bact. 54, 284.McClean, D. (1943). Biochem. J. 37, 169.Meyer, K. (1947). Physiol. Rev. 27, 335.

Hyaluronidase: Correlation between Biological Assayand other Methods of Assay

BY J. H. HUMPHREY AND R. JAQUESNational Institute for Medical Research, Mill Hill, London

(Received 25 June 1952)

Different workers have at various times sought tocorrelate the in vivo and in vitro activities ofhyaluronidase preparations. Chain & Duthie(1940), drawing attention to the association betweendiffusion in the tissues and the mucolytic activity oftesticular extracts, pointed out that the viscosity-reducing and diffusing activities of various enzymepreparations were in good agreement. McClean(1943) compared enzymes from different sources bytheir diffusing activity with Humphrey's (1943)technique, by their viscosity-reducing activity atpH 7*0 and by the mucin-clot prevention test; hefound that the results of the viscosimetric assaycould be very closely correlated to those of the skin-diffusing activity. Madinaveitia, Todd, Bacharach& Chance (1940), on the other hand, were unableto correlate satisfactorily the viscosity-reducingactivity of hyaluronidase preparation with theirskin-diffusing activity and stated that diffusingfactors ought not to be assayed in terms of theiractivity in the physicochemical test. In his reviewon the biological role of hyaluronic acid andhyaluronidase, Meyer (1947) pointed out that thespreading reaction could not be considered as anaccurate assay of hyaluronidase and that the corre-lation between spreading reaction and physico-chemical methods of hyaluronidase assay was poor.In certain instances the comparison was vitiated bythe use of different conditions of pH and salt con-centration in the different tests, but. even when thiswas avoided the error ofthe biological assays was toogreat to permit any conclusion that in vitro tests area true measure of in vivo activity. The question is ofsome importance because of the need to standardizehyaluronidase in view of its increasing use inclinical medicine. Having devised a biologicalmethod of hyaluronidase assay which gives re-producible results (Jaques, 1953) we undertook to

assay seven different enzyme preparations by themore convenient viscosimetric and turbidimetricmethods and to correlate the results with those ofthe bioassay. All tests were performed underapproximately physiological conditions of pH andsalt concentration in order to imitate more closelythe conditions ofthe skin test (McClean, 1941, 1943).

METHODS

Substrates. Two different hyaluronate preparations wereused. (a) A commercial potassium hyaluronate (Allen andHanbury) containing approximately 60% hyaluronate(measured as glucosamine) dissolved in 0-9% (w/v) NaClcontaining potassium phosphate buffer, pH 7 0, I =0-1 withrespect to phosphate, and 1:100 000 thiomersalate. Thispreparation will be referred to below as 'commercialhyaluronate'; it was used at 0-2 and 0 35% (w/v) in viscosi-metric and turbidimetric assays respectively. (b) Purifiedpotassium hyaluronate prepared by Dr H. J. Rogers fromumbilical cords by the method of Hadidian & Pirie (1948).This preparation contained no N other than glucosamine N.It will be referred to as 'purified hyaluronate'. It was used,in the same solvent, at 0 1 and 0-2% (w/v) in viscosimetricand turbidimetric assays respectively.The two batches ofpotassium hyaluronate were employed

for both the viscosimetric and turbidimetric assays of allseven enzyme preparations. In the light of the observationsof Meyer (1951) that various metals, notably Fe, are potentinhibitors of hyaluronidase, many of the assays withgelatin as stabilizer and commercial hyaluronate as sub-strate were repeated in the presence of 0-2% (w/v) sodiumpyrophosphate; this substance was added in all assays withthe purified substrate, although with our reagents it did notappear to have any appreciable effect.Enzyme solutions. These were made in 0900 (w/v) NaCl

containing phosphate buffer, pH 7-02, I=01 with respectto phosphate, containing 0-2% (w/v) gelatin and 1:100 000thiomersalate. Stock solutions containing 1-10 mg.enzyme/ml. lost no appreciable activity during 3 weeks'storage at 40.

J. H. HUMPHRRY AND R. JAQUES'Biological assay. The procedure was that described by

Jaques (1953).Viscosimetric method. The method of McClean & Hale

(1941) was used.Viscosimeters. The viscosimeters employed were of an

Ostwald type calibrated in absolute viscosity units over arange of viscosities studied. (Supplied by Dr L. A. Steiner,76 Cavendish Road, London, S.W. 12.) Flow time withwater at 300 was of the order of 65 sec. The water bath inwhich the viscosimeters were immersed was kept at 30± 10.The general conduct of the tests followed the lines suggestedby McClean & Hale (1941) who defined a viscosity-reducingunit as the amount of enzyme which will reduce the vis-cosity ofthe substrate in 20 min. to a level half-way betweenits original figure and that of the solvent employed. Ourresults were calculated on this basis.

Turbidimetric method. The method used was based uponthat of Tolksdorf, McCready, McCullagh & Schwenk (1949)in which the unit of activity is taken as the quantity ofenzyme required to reduce the turbidity given by 0-2 mg. ofpurified hyaluronate to that given by 0-1 mg. when incu-bated together for 30 min. at 370 in 0-1 M-acetate buffer,pH 6-0, containing 0-15M-NaCl. The turbidity is developed,after arresting the enzyme action by heating for 10 min. at600, by addition of a 1:10 dilution of serum or plasmapreviously heated at pH 3-1, and an excess of acetatebuffer, pH 4-2.

Since we wished to make comparison with other assaysperformed at physiological pH and salt concentrations, andto guard against inactivation of hyaluronidase which occursspontaneously in high dilutions in the absence of other

colloids, the method was modified. The buffered salinedescribed above was used throughout, and gelatin or, insome cases, gum acacia was added to give concentrations of0-1 or 0-25% respectively in the enzyme-substrate mixtures.The period of incubation at 370 was shortened to 10 min.,since evidence was obtained that the rate of enzyme actiondiminished quite markedly with further incubation.Furthermore, the reaction was stopped by chilling at 00,followed by addition of dilute acidified horse serum, ratherthan by heating at 600 which permits considerable furtherenzyme action to take place.

RESULTS

Table 1 compares the results of the three differentmethods of assay, the potency of the standardenzyme, which was always assayed simultaneously,being taken as unity. At least two assays, andoftenmany more, were performed with each enzymeby each method. The potencies of the variousenzymes as determined with the viscosimetricmethod are in reasonable agreement with those ofthe bioassay. In the viscosimetric assay the natureof the substrate apparently did not affect theestimation of potency of the majority of enzymestested, the only marked difference occurring withthe streptococcal enzyme (G) which was less activeupon the purified substrate. With the pure hyaluro-nate the value for G corresponds much better with

Table 1. Comparison of the activity of several hyaluronidase preparations using various assay methods

Bioassayactivity interms ofstandard

Enzymes (b)A, standard; testis 1B, testis 5-1C, testis 5-4D, testis 6-8E, testis 0-59F, Staphylococcus 2-3G, Streptococcus 0-53

Viscosimetric methodactivity in terms

of standard

Commercial Purifiedhyaluronate hyaluronate

(VI) (V2)1 14-5 5-46-6 6-97-6 7-21-0 0-612-6 1-71-4 0-65

Turbidimetric methodactivity in terms

of standard

Commercial Purifiedhyaluronate hyaluronate

(t1) (t2)1 12-1 2-22-6 2-33-8 2-60-5 0-380-8 0-540-2 0-16

Ratios

b/vl b/v2 blt1 bNt21 1 1 11-1 0-9 2-4 2-30-8 0-8 2-1 2-30-9 0-95 1-8 2-60-6 1-0 1-2 1-50-9 1-4 2-9 4-30-4 0-8 2-7 3-4

Table 2. The potency of several hyaluronidae preparationsConcentration of enzyme whichreduced the viscosity of the

substrate by half in 20 min. at 300

Substate Substratecommercial purifiedpotassium potassium

hyaluronate hyaluronate(0-2 mg./ml.) (0-1 mg./ml.)

5-5 7-031-19 1-30-83 1-010-73 0-985-4 11-52-04 4-134-05 10-9

Concentration of enzyme whichreduced the turbidity value of thesubstrate by half in 10 min. at 370

(,4g./ml.)Substratecommercialpotassiumhyaluronate

(0-35 mg./ml.)7134-52719-2

13391

370

Substratepurified

potassiumhyaluronate(0-2 mg./ml.)

9141-64035

240167570

EnzymesA, standardBCDEFa

60% I953

CORRELATION BETWEEN HYALURONIDASE ASSAYS

that obtained in the bioassay. The overall agree-ment of biological and viscosimetric assay is broughtout in the column wherein are listed the ratios ofboth assays with each substrate. On the other hand,the correlation between the turbidimetric resultsand this bioassay is less good. Thus, in the biologicalassay the preparations B and C were both about fivetimes as potent as the standard, whereas turbidi-metrically their activities in terms of the standardwere respectively 2-1 and 2-6. With the bacterialenzymes F and G the discrepancy between theresults of both methods was even more pronounced.In Table 2 some experimental values are given,expressed as reciprocal dilutions (w/v) of thevarious enzyme preparations required to reduce theviscosity due to the substrate by half in 20 min., orthe turbidity by half in 10 min., under the condi-tions outlined above.When 0-2 % (w/v) gelatin was used as stabilizer in

the diluting fluid, the shape ofthe curves ofturbidity(absorptiometer reading), after the set period ofincubation, against quantity of enzyme was closelysimilar for all the enzyme preparations, and therelative potencies calculated in terms of anyarbitrary turbidity end point were likewise similar.In earlier experiments 0-5 00 (w/v) gum acacia wasused instead of gelatin, and it was regularly ob-served that preparations A, B and D gave curveswhich had a steeper slope than those given by pre-parations C and G. The relative potencies in the gumacacia experiments were therefore dependent uponthe turbidity end point chosen, and consequentlyalso upon the time taken to reach this end point.

Table 3. Relative activities of enzyme preparationsper mg. compared with standard preparation by theturbidimetric method, using a single preparation ofthe substrate

(Method (a): incubated 10 min. in presence of 0-2%(w/v) gelatin. Method (b): incubated 30 min. in presence of0-5% (w/v) gum acacia.)

Enzymepreparation

BC

DG

Method

(a) (b)2-1 1-2626 2-53-8 1-402 027

The importance of this effect is shown in Table 3,where the apparent activities of four preparationsupon the same commercial hyaluronate substratecompared with the activity of the standard are

shown (a) measured in presence of 0-2 (w/v)gelatin with 10 min. incubation and (b) measured inpresence of 0-5 (w/v) gum acacia with 30 min.incubation. Whatever the cause of the discrepanciesin the presence of gum acacia, it appears to be un-

suitable for use as a stabilizing agent in this test.

DISCUSSION

The possession of a reasonably accurate biologicalassay method for hyaluronidase has made it possibleto test the validity of the correlation betweenactivity in vivo and in vitro, measured by viscosi-metric and turbidimetric methods. We have shownthat, provided suitable precautions are taken,among them the use of a reference standard enzymepreparation in each test, the correlation for theviscosimetric method is good and nearly as good forthe turbidimetric method. If the bacterial enzymesare excluded, which is probably justified since thereference standard was a testicular preparation, itmay be deduced that the viscosimetric assay gavean accurate measure of the biological activity,especially when the purified substrate was used. Theturbidimetric assays all appear to have givenrelatively low values compared with the bioassay.Inspection of the figures in Table 1, however, sug-gested that if one of the purified testicular enzymeshad been used as the reference standard, the resultsof the assays by all three methods would have shownratios close to unity, at least in so far as otherpurified testicular preparations were concerned. Thefigttres were therefore recalculated on the basis ofunit potency for preparation C, and in Table 4 areshown the ratios obtained. The agreement betweenthe results of the biological and the other assays isvery good for the purified testicular preparations,and it seems likely that for use as a practical bio-logical standard, valid for all three methods, a partlypurified testicular preparation would be preferableto crude seminal fluid, although the latter wasoriginally chosen with the idea that it would containany enzyme or enzymes to be found in the purerpreparations.The fact that the correlation is less good with the

two bacterial enzymes is not surprising, since thereis no evidence that they are single enzymes or thattheir action is identical with that of testicular pre-parations.The influence of the nature of the substrate upon

the results of in vitro methods of hyaluronidaseassay has been mentioned by many workers. Webegan this work with the expectation that if areference preparation of enzyme were used in allassays no influence of the substrate would beapparent, since all like enzyme preparations, atequivalent concentrations of active material, shouldbe similarly affected by variations in degree ofpolymerization, content of sulphated polysac-charide or presence of enzyme inhibitors. Thisexpectation was not fulfilled, and the apparentvariations in activity of the enzymes with the pureand the cruder substrate were outside the range ofexperimental error. Furthermore, the effect ofusing gum acacia as stabilizer instead of gelatin was

61

62 J. H. HUMPHREY AND R. JAQUES I953

Table 4. Ratio of activitims by different assay method8, when activity of preparation C 8 taken a8 1

Biological assay Biological assayViscosimetric assay Turbidimetric assay

Commercial Pure Commercial PureEnzyme source substrate substrate substrate substrateA, seminal fluid 1-26 1F26 0 5 045B, testis 1-38 1-22 1417 1-0C, testis 1 1 1 1D, testis 1.1 1P2 0-87 P1.E, testis 0.75 1-2 0 58 0 69F, Staphylococcus 0-91 0-58 0-72 0-540, Streptococcus 2-1 0-9 0-8 0-7

markedly different upon different enzyme pre-parations. These observations imply some degreeof heterogeneity of the enzyme preparationscompared. It is of interest, from a practicalpoint of view, that the closest agreement betweenin vitro and in vivo activity was obtained withthe purest substrate.

SUMMARY

1. Five preparations of testicular hyaluronidase,one of streptococcal and one of staphylococcalorigin were compared by an accurate skin-diffuiion

assay and by viscosimetric and turbidimetricmethods.

2. With testicular preparations the results ofthe three methods agree well, provided that thepotency is measured in terms ofa reference prepara-tion of enzyme, and that the pH and ionic strengthof the solvents are approximately physiological. Itis desirable to use highly purified substrate, and tostabilize the enzymes with gelatin rather than gumacacia.

3. The activities of the two bacterial enzymesshowed a close but not complete correlation betweenthe three tests.

REFERENCES

Chain, E. & Duthie, E. S. (1940). Brit. J. exp. Path. 21,324.

Hadidian, Z. & Pirie, N. W. (1948). Biochem. J. 42, 260.Humphrey, J. H. (1943). Biochem. J. 37, 177.Jaques, R. (1953). Biochem. J. 53, 56.McClean, D. (1941). J. Path. Bact. 58, 13.McClean, D. (1943). Biochem. J. 37, 169.

McClean, D. & Hale, C. W. (1941). Biochem. J. 35, 159.Madinaveitia, J., Todd, A. R., Bacharach, A. L. & Chance,M. R. A. (1940). Nature, Lond., 146, 197.

Meyer, K. (1947). Phy8iol. Rev. 27, 335.Meyer, K. (1951). J. biol. Chem. 188, 485.Tolksdorf, S., McCready, M. H., McCullagh, D. R. &

Schwenk, E. (1949). J. Lab. clin. Med. 84, 74.

The Differentiation of True and Pseudo Cholinesteraseby Organo-phosphorus Compounds

BY W. N. ALDRIDGEMedical Re8earch Council Unit for Research in Toxicology, Serum In8titute, (Jarshalton, Surrey

(Received 1 July 1952)

True cholinesterase is the enzyme present in thebrain and erythrocytes of many species whichhydrolyses acetylcholine at a higher rate thanbutyrylcholine and is inhibited by excess substrate.Pseudo cholinesterase is present in the serum ofmany species and hydrolyses butyrylcholine at ahigher rate than acetylcholine and is not inhibited byexcess substrate. Mendel, Mundell & Rudney (1943)introduced the substrates acetyl P-methylcholineand benzoylcholine for true and pseudo cholin-

esterases respectively. Differentiation of true andpseudo cholinesterases has also been achieved by theuse of selective inhibitors. Diisopropyl fluoro-phosphonate (DFP) inhibits pseudo cholinesteraseat a lower concentration than the true (Mazur &Bodansky, 1946; Mendel & Hawkins, 1947; Adams& Thompson, 1948). However, in some species thesensitivities of the two enzymes are not verydifferent (Ord & Thompson, 1950). A series oforgano-phosphorus compounds have been ex-