The Adjuvant Activity of Aliphatic Nitrogenous Bases - NCBI

ImmunologV, 1966, 11, 369.

The Adjuvant Activity of Aliphatic Nitrogenous Bases

D. GALL

Wellcome Research Laboratories, Beckenham, Kent

(Received 9th March 1966)

Summary. By the use ofdiphtheria toxoid in guinea-pigs, high adjuvant activity hasbeen found in a number ofaliphatic nitrogenous bases including amines, quaternaryammonium compounds, guanidines, benzamidines and thiouroniums. Activityappears to depend on a combination of basicity and a long aliphatic chain oftwelve or more carbon atoms. Such adjuvants tend to be haemolytic, and causedamage to tissue culture monolayers. It is suggested that their activity is connectedwith their surface activity and hence their ability to alter cell membranes, but thatthe basicity plays a further as yet undetermined role.

INTRODUCTION

The commonly recognized adjuvants are remarkable for their diversity; superficiallythere is little in common between alum precipitates, Freund emulsions, hexadecylamineand saponin. On closer examination, however, they all appear to be capable of bindingproteins, they cause local reactions at the site of injection, and are potentially or actuallysurface active.The perusal of a list of emulsifying agents (for example, in The Pharmaceutical Pocket

Book, Council of the Pharmaceutical Society of Great Britain, 1960) reveals a surprisingnumber of substances which at some time have been quoted in the literature as havingadjuvant properties. Apart from the more obvious examples, such as the surface activeagents used in Freund emulsions, methylcellulose (Rivenson, 1958), phospholipids (Weissand Dubos, 1956), alginate (Amies, 1959), saponin (Espinet, 1951; Johnson, Neal andGall, 1963; Richou, Jensen and Berlin, 1964), lanoline (Ramon, Richou and Staub,1937), quaternary ammonium compounds (Fulthorpe, unpublished observation) andcolloidal metallic hydroxides (Glenny, Pope, Waddington and Wallace, 1926, andnumerous others since) all appear as possible emulsifying agents. For this reason it seemedworth exploring the adjuvant activity of surface active agents generally, but as, with theexception of saponin, the activity seemed to lie mainly among the cationic quaternaryammonium compounds, and the most active of these were related to the amine hexadecyl-amine (Noll and Younger, 1959; Jensen and Woodhour, 1960), the emphasis of theinvestigation changed and the nitrogenous bases were explored at some length. Resultsobtained with saponins will be reported in a later communication.The regular occurrence of local lesions at the site of injection, the surface active nature

of some of the compounds (including saponin) and the presence of lipid soluble long alkylgroups in many of them suggested that there might be an action on the cell membraneassociated with adjuvant activity. Such an action was tested by their haemolytic activityon sheep red cells and by action on tissue culture monolayers.

369IMMUN D

MATERIALS AND METHODSAnimals

Albino guinea-pigs, 400-450 g, in groups of five, were used for most experiments.Albino mice, 18-22 g, were used for some confirmatory experiments.

AntigensA purified diphtheria toxoid of rather poor antigenicity, BJG 333, at 3500 Lf/ml, was

diluted in borate-succinate buffer at pH 7-5 to give a dose of 1 Lf in 0-2 ml. This dose wasgiven to the guinea-pigs, twice, at 28 days interval subcutaneously into the abdominalwall. Some experiments were done with a tetanus toxoid, TD 343, at 1200 Lf/ml, dilutedto 0 05 Lfin 0-2 ml in borate buffer at pH 8 1 and given in a similar scheme.

AdjuvantsIn all these experiments adjuvant was given only with the first dose of antigen.After some preliminary investigations, in which substances were tested in doses of 10 yug,

500 ug or 1 mg or over a range of doses, the dose of adjuvant was fixed at 100 pg. Formost substances this appeared to be enough to produce substantial adjuvant activity if anywas to be forthcoming, and in those substances in which the dose relationships were morecarefully studied it was found that if there was activity at this level, increasing the quantityof adjuvant did not materially alter the order of response. There are some exceptions tothis rule (notably amongst the silicas), and no doubt some activities have been missedthrough using too low a dose, but in general this seems unlikely to apply to the type ofcompound under consideration here.No attempt was made to combine antigen and adjuvant other than by simple mixing,

and no chemical coupling was attempted. After mixing, the solution or suspension wasgently agitated on a turntable for 1 hour before injection.Most of the substances tested were insoluble. Where necessary they were ground

to produce as satisfactory a suspension as possible, with occasional recourse to ultrasonicdispersion. Very often the suspensions were crude, with a fair range of particle size, but aslong as the suspensions were injectable this was disregarded. With hexadecylamine, forexample, the coarsest suspensions seem to be just as active as the finest. In the early days ofthe investigation, and especially with the insoluble liquids, suspensions were sometimesmade with the aid of 1 per cent Tween 80. Recently, dimethylformamide has occasionallyproved useful.

Estimation of antibodyGuinea-pigs were bled routinely by cardiac puncture 10 days after the second injection.

In some experiments they were also bled just before the second injection. Diphtheriaantitoxin titres were measured by the guinea-pig intracutaneous method, and tetanusantitoxin by the mouse subcutaneous method, at approximately 100 per cent differences.

In the tables the results are expressed as the number of individual animals reachingtitres at ten-fold intervals, together with a 'score', and the geometric mean titre wherevalid. The score is equal to the log geometric mean titre plus 4 (to make all values positive),and makes the arbitrary assumption that animals with titres less than 0-001 u/ml have atitre of 0-0001 u/ml. The score is a convenience to enable test groups to be ranked eventhough some animals have not reached a measurable titre and a geometric mean wouldtherefore be invalid.

370 D. Gall

Adjuvant Activity of Aliphatic Nitrogenous Bases

Much of the reported work on adjuvants has been done on the primary antibodyresponse. The secondary response was adopted here because it was found to give a widerrange of difference between good and poor responses than did the primary. With theadjuvant given only with the first dose of antigen, the second dose takes on the nature of achallenge, and the secondary response becomes a measure of the degree of basal immunityinduced by the adjuvant rather than of the amount of antibody production it will elicit.The dose of diphtheria toxoid chosen gave, without adjuvant, mean responses well

below 0 001 u/ml (Table 1). With a good adjuvant the responses were near the maximumthat can be expected of a two-dose method in guinea-pigs (10-50 u/ml).Estimation of haemolytic action

Fresh sheep red cell suspensions were diluted in saline to contain approximately 60,000red cells per ml. Substances to be tested were dissolved or suspended in saline at twice thefinal concentrations of 102, 5x 10-', 2x 10-', etc., to 10' or more. Volumes of 0.5 mlwere mixed in the cups of transparent plastic trays and allowed to stand, covered, for24 hours at room temperature.

Haemolytic activity was judged by viewing ruled lines on white paper through the cupsand was recorded as the negative log concentration at which total, partial and tracehaemolysis occurred.

Tissue culture monolayersMonkey kidney or HEp 2 monolayers were used. Routinely, HEp 2 were preferred as

they gave a more uniform and reliable cell sheet. The results were essentially similar witheither type of cell.The cultures were used 24 hours after seeding into test tubes containing 3 ml of 199

medium, without serum, covering an inclined coverslip. The serum was omitted as it wasfound to inhibit the changes brought about by some of the substances, and its absence didnot affect good growth over a short period. A volume of 0 1 ml of solution or suspension ofthe substance to be tested was added to give the same final concentration (0-5 mg/ml) aswas used for injection with toxoid into guinea-pigs. All samples were put up in duplicate,with control tubes which were left undisturbed. The cultures were read after a further24 hours incubation. The monolayers were fixed and stained (in Helly fixative followed byLeishman stain) without removing the coverslips from their tubes. They were thenextracted and mounted on slides for examination.Two controls were examined at the time of addition of the test substances and two more

at the end of the further incubation. The monolayers were examined microscopically andany changes in the cells recorded.

RESULTS

GENERAL LIMITS OF SIGNIFICANCE

Because only small groups of animals were used for each test statistical analysis of theresults has not been attempted. Although titres were tested to approximately two-folddifferences they have been grouped in the tables at ten-fold differences and thereforeminor differences between groups have been smoothed out to some extent.

Conclusions can only safely be drawn from major differences between groups. For themost part, where tests were -repeated they repeated satisfactorily, that is to say highly

371

active substances repeated as highly active and moderately active as moderately active.Weak activity was more variable in its reproducibility. In a few instances there werefailures to repeat satisfactorily and this generally shows up as a wide scatter in the responseswhere more than five animals have been used. In the middle and low ranges of responsethe scatter is always greatest and differences between means least reliable. In general,ten-fold differences in mean titre, or differences of l 0 in score, are probably significant.In this paper conclusions will, on the whole, be based on much larger differences.The present scheme has been adopted in order to separate substances with adjuvant

activity from those without activity. It has not proved sensitive enough to distinguishclearly the better oftwo good adjuvants.

Soon after the work started some changes were made in the dietary regime of theguinea-pigs. Recent observations in these laboratories suggest that such changes can affectimmunological responses to some extent, but although they may have contributed toscatter in the bulked responses of repeated tests they do not affect the order of response.

CONTROLS

The titres reached by guinea-pigs receiving toxoid without adjuvant are shown inTable 1. Seventy-two per cent of the animals failed to reach a titre of0001 u/ml and only3 per cent reached 0 1 u/ml.

TABLE 1

CONTROLS: TITRES REACHED BY GUINEA-PIGS RECEIVING TOXOID ALONE

No. of guinea-pigs reaching titres:No.

Dose animals 3 0 001 0 01 01 1.0 10 Score< 0-001 0*01 0.1 1.0 10 100

Diphtheria toxoid,purified, BJG 333/3, 1 Lf 68 49 12 5 2 0-46alone

ADJUVANT ACTIVITY OF SURFACE ACTIVE AGENTS

Several of the non-ionic surface active agents showed slight activity (Table 2), notablySpan and Tween 80. The anionic agents appeared to be inactive. Tween 80 was quitehighly active when used neat as the toxoid diluent, but the degree of activity in dilutesolution was completely overshadowed by that of the cationic quaternary ammoniumcompounds (Table 3).Among the quaternary ammonium salts the influence of the long alkyl chain stands out

prominently, in spite of some exceptions. Where a long alkyl chain is absent little or no

activity can be expected. Where a long alkyl chain is present, hexadecyl, octadecyl andlonger chains tend to be more active than their shorter counterparts. Among the tri-methyl ammonium salts (Table 3, a) the tetradecyl compound shows higher activity thanwould be expected and two of the hexadecyls lower activity, but it must be rememberedthat many of the commercially available preparations, of which these are examples,consist of mixtures of various alkyl quaternaries depending on the natural fat or oil fromwhich they have been prepared.

D. Gall372


000,~c c000,-d "tC4 q0 0 0! 0000C4r-~ N-C40 (00LO - M -LO 01QOC)

- CM~~~~~~~~~-

--4 CM4 - -C)

1 C

CO L 0 O L LO LO LO LO LO LO Id trl- LO LO) LO LO) LO)

.____ _0000.*-*- .

6 oo 66 o o 6 :__ : :

0

0 0 ~~~~0

C'SC's 0 I-

- v

Ucd ~ ~

0~~~~~

-0

4)0C)c,

373

4)* .0

4)bo._

CUCI

4)._

bo

0

6z.

o

*. 0

0-

-.

o o5--I

000

/A

0

6CZv):-r.*

3

CQ

zo0

uwu-t

C4

0

z

0

z04 z

0

_ _- _- -_

LO I' LO LO LO LO

374 D. Gall

tt*O .Lt~(0Mo 1S ? C_0 C- NC9C 00tO- D0 0

sQ cs 00oONta n a) -C X- LO 0 00dCD O o o oC00co 00OCM

08 eo--C4- c --o - co

-CM 00 CO c0 C 1- CM4N CO (0 -(0-CM

NtO _'

O~~~~~~- CO Oe _IIC1 cc4~e e

-4 -CMIC

0

-~~~~ CM~-OC0 -I- CI4CMC ( - M C C Mn

O --=e -oo CM-CM - - - _ I -

0z oC80A v

cly

z E to LO. LO LO° Looo Lo o ) o OOCO0o L LoLo-o o o oo o o o o o _

0

66666666 6 6 66 6 6A'66666 66Az

ce 0 cl~~~~~~~~~~~~~~~~~~~~~~~~~~

0> 0o 0 ~ ~,, ~ ~ ~ ~ > ~ ~ ~ C~l C~l C~l N C~l C~l N CEW 4:N < ) s) CO0CI4001 -_

z O~~~~~~~~~~~~~~~~~~~~~~o <0 S ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~M.-.e0

Z 0 ~ ~ ~ ~~'0 0 oS.0o~~~~~~~~0 0 0 L

0~~~~~ ~ ~ o~~~~~~SS - ~~~~~~~~~~~~0got~,3.- r 00m0'- 0 C's00%00 CZ-

4) -0 )5. 0* 0

Oc005 0voo 0Z

0 1-4

O~~~~~~~~~~

0 00 O :o Q~.00 00 S 5 0

0~~~~~0o:3 ::3 :Z Cd C.00~5oZ N

00.000.fl0,~~~~ *- 0 = 0 5 0055 ~

E E E 0SC's 0~~~ t~0 E E -OXct~50

-g~~

- 0 S5~~~~~~ 0 E X 71~ >

000d 4-,i Ur0- >4 0 - 0"t3 Q.) .s.AE E '

C.) C.) C 4 0-x< x O AM ''CZCMeMMM)MMMCC0C~CC

I

I

i

Adjuvant Activity of Aliphatic Nitrogenous Bases 375

Among the alkyl benzyl dimethyls and the dialkyl dimethyls (Table 3, b and c) thesame increasing activity with increasing chain length holds good. Among the seventeenmiscellaneous compounds five have alkyl groups of twelve or more carbon atoms; ofthese, three show high activity (17, 21 and 28) and two moderate activity (22 and 33).Of the remaining twelve only one (20) is active and this contains a tetramethylbutylgroup.The presence of the benzyl group, in the alkyl benzyl dimethyl compounds, appears to

lower activity somewhat. This seems to be a general finding among the aliphatic nitro-genous bases. On the other hand, the dialkyl quaternaries produce slightly higher titresthan their monoalkyl counterparts.

ADJUVANT ACTIVITY OF AMINES AND AMINO COMPOUNDS (Table 4)

As with the quaternary ammonium compounds, it is the amines with the long alkylgroups which are the most active. Hexadecylamine has been extensively reported on as anadjuvant by Noll and Younger (1959), Woodhour, Jensen and Warren (1960), andYounger and Axelrod (1964). In this series it serves as a convenient reference substance,and it will be noted that only six amines tested give comparable or better responses:octadecylamine (8, 9), hexadecyl- and dihexadecylmethylamine (12 and 17), dioctadecyl-amine (19), hexadecylethylenediamine (27) and 12,21-diaminodotriacontane (29).Although, considered on its own, octadecylamine does not give significantly betterresponses than hexadecylamine, the fact that dioctadecylmethylamine is also slightlysuperior to dihexadecylmethylamine suggests that the octadecyl compounds are slightlysuperior to the hexadecyl.

Unfortunately there are many gaps in the series yet to be filled, especially among thelower alkyl amines. In addition, the short chain amines are often immiscible oils whichare difficult to prepare in even dispersions; hence the rather extensive use of Tween 80.Although Tween 80 appears to have little influence on the activity of highly activeadjuvants like hexadecylamine, it probably does raise the activity of less active compoundsand has probably exaggerated the activity of the lower molecular weight amines.Among the long chain alkylamines there are some interesting discrepancies. In par-

ticular, dihexadecylamine (13), in comparison both with the dialkylmethylamines andwith its quaternary analogues, gives surprisingly low results. On the other hand thedialkylmethylamines give high responses in comparison with the alkyldimethylamines.The contrast in activity between 5,14-bisdimethylamino octadecane (28) and 12,21-

diaminodotriacontane (29) is also interesting and may be related to the methylation of theamino groups.A number of miscellaneous amino compounds tested included the following:

UreaBiuretGlycineLysinea, 8-Diaminopimelic acidEDTA disodium salt5-Chloropentyl carbamateMethyl carbamoyl acetic acid hydrazidet-Butylaminoacetonitrile hydrochloride

LO LO LO o LO U o) LO O LO LO LO LO LO LO LO o LO LO LO LO LO

- - - - - -

oo

u ~~~C)"0

c

00

-- c

H~ ~ ~ ~ ~ ~ o H o-H-

sZ 0 vo

O-CM X + ¢ Q t-00 m O- X + e1-C00 O O-CM

~ -0 CZ 0 Z v V

s_~~~~~~~~~~~~~~- 0

D. Gall

N I T £ cpcN 1 CN- 0-,rf)

c LO 1c4 oo 00 0LO MC14 ' Z m lZ 'd o CO t- 00 1.1 cn 0* . . * . . . . . .

. -.I

00-0-~~~~~~~~~~ -4-I N o? IoI.- P _ _ _7 o,C14 _- co co o o 4 Cj o

00o )o " 0 0)C* *

1--d O c

q " c 4 en ooC'4 - 4 0) C4 dtC1 -

00o' C 00C -0C_ 40o0oo0 CN m lt Co CN CO CN Ca00OC -. C. . &

c14 _- n co (£ Ul c14

_-4 1-4--4 I" C')C- C' ) - C'4C'4- eN - - -

eC - CN

- -

- - eN

I - -.4 - cq

C')eN -

LO LO I'd CN LO LO o LO LO

,:,n,:- -tj- u:

Lo Lo Lo LO - -Lo

-, . . . . .6. .0 .c_3

I--

I N

N I) N - - -s

LO - C' - ' -t C-4 Lo

Cl)

o 0

000

0-

_0_

0

0

l:o

:z

376

00zz0

U

ci0z

2

z

I'l

u

v4.1

4.1

bir.xum(U

Wbicl

4.)0zbf

L6..0

6-- C') co -I-

Adjuvant Activity of Aliphatic Nitrogenous Bases 377

enc14 ce) C1 f LO1Q

CNO ~~~~~~~~~~0 C~~~~~~~~ liii I Il ~0Ll TC~L~ ?I II Iq I4LsOs O O W ) LoI,IqI I, - I- I4 I- c-O 0d604

0 ~ cCqCN 00L0 0C400l

004 0004.O 00000000d 0n C O OOr00mr-c) O

co t-. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

-ce a) ce CO LO - C'l- C-4

4- CN N ,d LO Lo .Id Lo CM m m t M cN

ce)cnC --4 Ce' CN coCN4 Cn CO I- - CN _

~--

C-4- -

00 LO 0 LO LO U-) LO) 000 LOoooo~~~~~nenooox-4C'

to b

owUF o.$,VwV V 0

C

z z 0; a: z z 05,55555555555

'............-e31 1 1 1 1-01 O-C1

c-4_ _ _

LO LO LO) LO U LO) LO LO LO) L LO 'I, 000 LO LO LO) 0 LO U-) 0 tr U-) LO LO

. . . . . . . . . . . . . . . . . . . .

-ce) " LO (. r- 00 C)CO- N en ,d L C. r- 6oo~ 6O-: ce; 4 Lei ( r 00.AL - - - - - - - C' C1 C14 4 CN C' C4 C' CN C'4 e ne n n CO CO CO CO CO

* C0

c)

UO

bO0

/A V

0

IA

&O. C,0E

V -

0 E

z

z

0

v

~zz0z

4II.z

t-E.c0a;.

rot

CI4

I

tf..c

c

1- (1 O-Aminodecyl)-4-methylpiperazine trihydrochlorideDodecyl hydrazine hydrochlorideDodecanoic acid hydrazide1 -Amino-2-hexadecyloxyethane hydrochloride1-Hexadecylamino-2-hexadecyloxyethanel-Ethoxycarbonyl-4-hexadecylaminothiocarbonyl piperazineHexadecanoic acid hydrazideHexadecylacetamideEthyl N-hexadecylcarbamateHexadecylisothiocyanateOctadecamide1 0-KetooctadecamideOctadecanoic acid anilide

Of these, only hexadecyl isothiocyanate (in admixture, not chemically coupled) gave amean titre above 0*01 u/ml (score 2*86). 1-Amino-2-hexadecyloxyethane hydrochloridegave variable results, with a few animals over 1 u/ml, but none of the remainder, eventhose with long alkyl chains, showed significant activity.The mere presence of a long alkyl chain is evidently not enough. Long alkyl amides,

acetamides, hydrazides and anilides do not appear to be active.

GUANIDINES (Table 5)

The relatively complete series of dimethyl guanidines available (Table 5, a) provided amuch clearer indication of the influence of alkyl chain length. Activity began at aroundten carbon atoms, climbing steeply between eight and twelve carbon atoms and thereafterrising only slowly. As with the amines and quaternaries, the octadecyl compounds areslightly more active than the hexadecyl. As a group the guanidines appear to be aspotentially active as the amines.The same picture is reflected in the miscellaneous guanidines (Table 5, b). Those with

short chains or ring structures only are uniformly inactive, but those with alkyl chains oftwelve or more carbon atoms are all active.

It is interesting that among the guanidines the degree of methyl substitution appears tohave relatively little effect. There is some falling off of titre through the series: hexadecylguanidine (26), hexadecyl dimethyl guanidine (9, 10), hexadecyl trimethyl guanidine(27) and hexadecyl methyltetramethylene guanidine (30), but the effect is slight and thedifferences of doubtful significance.The two representatives containing two long alkyl chains, didodecylmethylguanidine

(25) and dihexadecylmethylguanidine (34), produce titres as high as their monoalkylcounterparts.

BENZAMIDINES (Table 6)

There is a suggestion of activity from a number of the benzamidines, especially thediethylaminoalkyloxy-benzamidines. The only two showing reliably high activity, how-ever, are benzamidines with a long alkyl chain, diethylaminododecyl benzamidine (12)and hexadecyl benzamidine (13).

378 D. Gall


O Oentos~~~~~~~to++csb-s~-4'IoU 0Q..............oV)~ 'OOOO-C t--+.OOooooS&4 A _

00

*O 00L MC C~4 LO

A V

Cd

Lo Lo LoL L oLoULoLoI oL

W2wz

Q

a

Nz

mm

5Ev4

¢i

4 . -r

0,o

(:E

379

D. Gall

0 -c cDoo o CDQ0CCD. . . .0 .) . .i . .

C,

66666666666oo

Cs - _- Ce4 0

-0 en ooo= 0o-.0 0

. . ..

.C14

C& 0 O 04O C& co

0000 1C00 coo0000o o o .o0( CM* . . . * . *) . .:

1-4 CM C - CCM

CM CM - --- 1-

LOn CM4 co

- I I'*- I -

- I CC) I- -

CM

c _-

0 0) 0 m CM d LO LO

000LOOO LO LO LO L O4- -4 --

0

U sU

<) Z -ZO

_ ~z0~~0

i

_/

0 LO 000 LO L 0- - CM4--4-- 1-LO Ln L LO LO LO LO LO LO

0

P4.4 04Er, ;?.

uz4P4

U UC'

a .c

t

£EEE<

.o UU.F H t t w

g~~~~~~~~~~~~~1:24 C1:<q C4 C1:2N C1CN^ :^:

cn

C)

0c)

boRo o

0

-o

0)

0 o6 0

z o~

IAVJ

380

wxz

0

0

N

z

z

N

U)

N

CU

H

z0

N

NW4

0

Hw

W

0

H

U~H

z

¢-cn 4

ci


THIOUREAS, THIOSEMICARBAZIDES, THIOSEMICARBAZONES AND THIOURONIUMS (Table 7)

None of the thioureas, thiosemicarbazides or thiosemicarbazones showed adjuvantactivity, whether or not a long alkyl was chain present. These compounds, bearing the= S group, are neutral or near neutral. In contrast, the thiouroniums, which are stronglybasic, showed considerable activity when a long alkyl chain was present, even when thisoccupies the central part of the molecule, as in the bis-thiouroniums.

There were, however, exceptions. N-hexadecyl-S,N',N'-trimethyl thiouronium (13)is less active than the dimethyl (12), which may relate to the previous observations onincreasing methyl substitution. Similarly, the dihexadecyl methylthiouronium (14) wasinactive, in contrast to the monoalkyl compound (12). The S-hexadecyl-NN'-dibenzyl-thiouronium (16) was inactive compared with the S-hexadecyl-NN'-dimethyl (15), and itmay be significant that the least active of the long alkyl bis-thiouroniums, eicosamethylenebis-(benzylmethylthiouronium) (28), is the only one to contain a benzyl group.

MISCELLANEOUS LONG ALKYL COMPOUNDS

A few mainly non-nitrogenous long alkyl compounds were also tested:Octadecyl aldehydeOctadecanoic (stearic) acidDL-cf-Octadecylglyceryl etherGlyceryl- 1,3-distearate-2-azelateLecithin (egg)Octadecyl diphenyl phosphateOctadecyl mercaptan

The activity of these compounds was uniformly low and reinforces the observation thatlong alkyl compounds per se are not adjuvant.

RESULTS WITH TETANUS TOXOID (Table 8)

The adjuvant activity of a limited number of compounds was confirmed with tetanustoxoid. But while hexadecylamine and octadecylamine, the dialkyl quaternary Arquad2HT, and the thiouroniums gave high responses, the guanidines showed only moderateactivity with considerable scatter in the individual responses.

RESULTS IN MICE

The same limited number of compounds, with the addition of octadecyl trimethylammonium chloride (Arquad 18), were tested in mice with both diphtheria and tetanustoxoids.Mouse responses were generally poorer than guinea-pig responses, as is usually the case.

Only one compound, Arquad 2HT, gave consistently good results in the mice. Withdiphtheria toxoid, Arquad 18, hexadecyl and octadecyl amines, and the octadecylguanidine and thiouronium produced moderate responses, the hexadecyl guanidine andthiouronium poor responses. With tetanus, moderate activity was shown by the aminesand by octadecyl thiouronium, doubtful activity by Arquad 18, and none by hexadecylthiouronium or the guanidines.

381

d. LOen -AeN C4 _- C14

o CDo C,4 LoCqn C)O -l-C(C)<cq

O~ 6 LnOL) l~C

L oCq CqLo

-CM-

- c..LO

_ _

0~~~~~~~~1

c w.l 0 0 0n EEEEEEEN

O --- _ _ _ _ _

. . . . ..V

0;0

0

382 D. Gall

V -

O .E

U

0z

O00

-04

0-4

/A

bl)

~0r.X

$.

0A

z4)

*0

0

0x

Ez

F-

0

z

E:10

0

F-co

5


LOCAL LESIONS AT THE SITE OF INJECTION

The site of injection was inspected routinely on the 3rd day after injection and anestimate made of the lesion diameter. It soon became evident that a palpable local lesionwas very suggestive of adjuvant activity. Active substances nearly always produced apersistent local reaction in the tissues, generally a palpable nodule of 3-4 mm or more, andhistological examination confirmed that these were cellular infiltrations. The greatmajority of inactive substances caused little or no local reaction. There were exceptionsin both directions, but fewer adjuvant substances were unaccompanied by local reactionsthan the reverse. There was, however, only a very rough correlation between the size ofthe local lesion and the height of the immune response. But it was clear that, whateverreaction there may be between adjuvant and antigen, there was also an action of theadjuvant upon the recipient's tissues.

HAEMOLYTIC ACTIVITY OF ADJUVANT SUBSTANCES

The results were far from clear cut. Although most of the substances showing adjuvantactivity were also haemolytic there were several haemolytic substances which were notadjuvant, which correlates with observations on the surface active agents generally.The degree of haemolytic activity, as measured by the minimal haemolytic concentration,did not correlate with the degree of adjuvant activity. It is more helpful to compare theadjuvant properties of substances which show any haemolytic activity with those ofsubstances showing none.Among the non-ionic and anionic surface active agents there was least correlation. Of

seven agents tested, none was significantly adjuvant, but five were highly haemolytic andone moderately so.

Thirteen of the quaternary ammonium compounds were tested; eleven were stronglyhaemolytic and two were without haemolytic action. The eleven haemolytic compoundsconsisted of eight which were highly adjuvant, but three which had only weak activity.The two non-haemolytic compounds were trimethyl ammonium acetohydrazide chloride,which is inactive, and the highly adjuvant dioctadecyl dimethyl ammonium iodide. Thisresult has been repeated and confirmed.Among the amines the results were slightly more consistent. Of twelve compounds seven

were haemolytic and five non-haemolytic. Of the seven haemolytic members four werestrongly adjuvant, two erratically adjuvant (dimethyl hexadecylamine and dimethyloctadecylamine) and one inactive (5,14-bisdimethylaminooctadecane). The five non-haemolytic members included one highly adjuvant substance (12,2 1-diaminodotri-acontane), one moderately active (hexadecylisothiocyanate) and three inactive.The dimethyl guanidines, arranged in order of increasing alkyl chain length, gave

coherent results. The decyl and higher alkyl members were all haemolytic and, with theexception of the decyl compound which gave erratic results, were strongly adjuvant.The three lower members of the series were neither haemolytic nor adjuvant.The sulphur-containing compounds have not yet been studied in any number. The

two active thiouroniums, S-hexadecyl-NN'-dimethyl and N-octadecyl-SN'-dimethyl,are both haemolytic.

ACTION ON TISSUE CULTURE MONOLAYERS

In general, there was a much more consistent action by adjuvant substances on tissueculture cells than there was on red blood cells. The observable action varied from mild

383

changes in the cytoplasm, producing, for example, a 'felted' or granular appearancewithout much other change in morphology, through more gross changes in both cytoplasmand cell outline, described as 'cobwebby', 'stringy' or 'coagulated', to extreme shrinkage ofthe cells with pyknosis and dissolution ofthe entire cell sheet apart from a few pyknoticindividuals. The type of change produced was quite characteristic of the class of substancebeing tested, and after a little experience it was possible to say from the appearance of thestained monolayer what type of compound had been acting on it.Three non-ionic surface active agents, Arlacel A, Tween 80 and polyethylene glycol,

produced very little change except for 'felting' of the cytoplasm. Ammonium stearate,aluminium stearate, ethyl stearate and glyceryl monostearate had no effect on the cellsin the 24-hour test period. Lissapol N, on the other hand, regularly caused completedissolution of the cell sheet within a few hours.

Eleven of the quaternary ammonium compounds were tested. Ten of these producedtypical changes, a stringy, coagulated appearance of the cells without disruption of the cellsheet (Fig. 1). Of these ten, sevenwere strongly adjuvant and three were erratically adjuvant(low scores, but high titres from a few individuals), but again there was no correlationbetween the degree of cell destruction and adjuvant activity. The remaining compound,benzoyldimethylphenoxyethyl trimethyl ammonium methane sulphonate, was inactiveboth against the cells and as an adjuvant.The amines also produced a characteristic change, very similar to the quaternaries,

coagulative, but less 'stringy' and often progressing to dissolution of the cell sheet. Ofnineteen amines tested, six produced complete dissolution, five typical amine changeswith or without some disruption of the cell sheet, and eight no significant change. Of theeight producing no change none was adjuvant. Of the five producing moderate changeone was not significantly adjuvant (diaminodecane) and one was the erratic l-amino-2hexadecyloxyethane. Of the six causing dissolution of the cell sheet only one, dodecylhydrazine, was inactive as an adjuvant. The remainder, hexadecylamine (Fig. 2),octadecylamine, hexadecylmethylamine, dihexadecylmethylamine and dioctadecyl-methylamine, were all strongly adjuvant.The guanidines tended to produce a more granular, or fine cobwebby, appearance,

sometimes progressing to complete destruction. The dimethyl guanidine series againshowed no activity among the three shortest members, and progressively more severeactivity amongst the rest, with the exception that the decyl compound caused much moresevere damage than its place in the series would suggest. The action of the hexadecylcompound is illustrated in Figs. 3 and 4. Four other non-adjuvant and two other adjuvantguanidines were tested. The non-adjuvant were all inactive against the monolayers, thetwo adjuvant members caused typical guanidine changes.

DISCUSSION

Any general conclusions that can be drawn from the foregoing results should properlybe confined to diphtheria toxoid in guinea-pigs. Experiments with tetanus, and in mice,have to some extent confirmed the results, but they have not always done so, and theranked order of activity is different. Furthermore, while the particular toxoid chosen(toxoided in the presence of glycine) was quite sensitive to enhancement, a toxoid ofsimilar flocculating power but considerably higher antigenicity, toxoided in the presenceof lysine (Linggood, Stevens, Fulthorpe, Woiwod and Pope, 1963), proved difficult to

384 D. Gall


iW''';; ''M' -~~~~:"Lsoto ...> °.:

3'

tH ....

1.eV.:.;.. }wwrv l_

=.: ^

- 4l.. r

°- SgL:...1_-B-L

}L:.:... :9.. -r:.

::. .: tt wi . .. }.: 5:

:

:iO.

- w::.

*- L

-

.- :w* : ;s - :. :w :.: w.

FIG. 1. Monkey kidney monolayer treated with dialkyl quaternary ammonium compound, Arquad2HT, 0 5 mg/ml, 24 hours. Leishman. x 420.

FIG. 2. Monkey monolayer treated with hexadecylamine, 0 5 mg/ml, 24 hours. Leishman, x 420.

(Facing p. 384)

` O."

iqw:.-- 40

A:

AI-

It

[,

D. Gall

3q

A.5~.~

..

FIG. 3. Normal HEp 2 monolayer. Leishman, x 420.

FIG. 4. HEp 2 monolayer treated with hexadecylguanidine, 0-5 mg/ml, 24 hours. Leishman, x 420.

.4- V


enhance further. Experiments performed by colleagues with viral, clostridial or protozoalantigens, with these and other adjuvants, emphasize the dangers of arguing from oneantigenic system to another, although some degree of adjuvant activity is often manifestwith more than one antigen. It seems as though adjuvants have potentially a broadspectrum of activity but may fail with certain antigens or in certain species for a variety ofreasons.

In some respects, the requirements for adjuvant activity among the aliphatic nitro-genous compounds emerge as fairly well defined: the compound must be basic and mustcontain a long aliphatic chain. Neither characteristic is sufficient alone, but within theserequirements a wide range of compounds is active.

Such compounds, having strongly polar and non-polar parts to the molecule, arepotentially surface active. With a few exceptions they caused disturbance of the tissuesat the site of injection, disrupted the cytoplasm of cells in culture and were lytic to redcells, which strongly suggests that adjuvant activity is associated with an action on hostcells. In the present series however, there was no correlation between the severity of theaction on cells and adjuvant activity, and both Freund emulsions and saponin tend toproduce much larger local granulomata than, for example, hexadecylamine or aluminiumhydroxide for a given adjuvant effect. Nevertheless, such behaviour recalls the action ofsaponin, also adjuvant, in altering the ultrastructure of cell membranes (Dourmashkin,Dougherty and Harris, 1962; Muir, 1962; Husson and Luzzati, 1963). Colloidal silicaalso, some preparations of which are adjuvant (unpublished observations), has beenobserved by ourselves and by Harley and Margolis (1961) to cause haemolysis or agglutina-tion of red cells. It is tempting to think of the adjuvant opening up pores in the lipo-protein mosaic of the cell membrane and allowing the antigen to enter the cell morereadily, either by a process akin to the 'appositional' phagocytosis of Carr (1962) or the'piggy-back' phagocytosis of Sbarra, Shirley and Bardawil (1962).However, Bangham, Rees and Schotlander (1962) have shown that it is basic long-chain

compounds which are the most effective in protecting liver cells against damage by carbontetrachloride, and on the basis of protection against leakage of /3-glucuronidase, havesuggested that they cause a decrease in cell membrane permeability. Subsequent work onthe action of saponin on membranes and cholesterol films (Bangham and Horne, 1962;Glauert, Dingle and Lucy, 1962; Bertaud and Hedley, 1963) has shown that the effectis more than a simple one of punching holes in a pre-existing mosaic.

Apart from alkyl chain length, the arrangement of the alkyl groups around the nitro-genous base also influences activity. In general, dialkyl compounds are as good as, if notbetter than, the equivalent monoalkyl, but two marked exceptions are di-hexadecylamine and di-hexadecyl methyl thiouronium. The former is in unexpected contrast toboth the monoalkyl amine and methylamine and to the di-alkyl methylamine, as well asto its quaternary and guanidine analogues. The adverse effect of increasing methylsubstitution has also been mentioned. These observations at present only underline thecomplexity of the phenomenon. Possibly more detailed physico-chemical characterizationof such compounds, together with studies of their effects on natural and artificial mem-branes, would shed further light on the nature of adjuvant action.Nothing has been said of the ability of these compounds to absorb the antigen and act

as a depot for slow release, or as a carrier to facilitate entry into cells. Some preliminaryattempts to estimate adsorptive capacity by the method of total combining power havenot been very successful with these compounds and this aspect has yet to be investigated.

385

386 D. Gall

ACKNOWLEDGMENTS

I am indebted to most of my colleagues for their helpful discussions and suggestionsat one time or another, and I should like to thank in particular Miss Mollie Barr and DrC. G. Pope. Mr Edwin Potter has ably carried out much of the work ofpreparing, injectingand bleeding.

REFERENCES

AMIES, C. R. (1959). 'The use of topically formed cal-cium alginate as a depot substance in active im-munisation.' 7. Path. Bact., 77, 435.

BANGHAM, A. D. and HORNE, R. W. (1962). 'Action ofsaponin on biological cell membranes.' Nature(Lond.), 196, 952.

BANGHAM, A. D., REES, K. R. and SCHOTLANDER, V.(1962). 'Penetration of lipid films by compoundspreventing liver necrosis in rats.' Nature (Lond.), 193,754.

BERTAUD, W. S. and HEDLEY, R. H. (1963). 'Hexa-gonal patterns in cell membranes.' Nature (Lond.),200, 89.

CARR, I. (1962). 'Appositional phagocytosis.'_7. Path.Bact., 83, 443.

DOURMASHKIN, R. R., DOUGHERTY, R. M. and HARRIS,R. J. C. (1962). 'Electron microscopic observationson Rous sarcoma virus and cell membranes.' Nature(Lond.), 194, 1116.

EsPINET, R. G. (1951). 'Nuevo tipo de vacuna anti-aftosa a complejo gluco-virico.' Gac. vet. (B. Aires),13, 265.

GLAUERT, A. M., DINGLE,J. T. and LucY,J. A. (1962).'Action of saponin on biological cell membranes.'Nature (Lond.), 196, 953.

GLENNY, A. T., POPE, C. G., WADDINGTON, H. andWALLACE, U. (1926). 'Immunological notes.XVII-XXIV.' J. Path. Bact., 29, 31.

HARLEY, J. D. and MARGOLIS, J. (1961). 'Haemolyticactivity of colloidal silica.' Nature (Lond.), 189, 1010.

HUSSON, F. and LUZZATI, V. (1963). 'Structure of red-cell ghosts and the effect of saponin treatment.'Nature (Lond.), 197, 822.

JENSEN, K. E. and WOODHOUR, A. F. (1960). 'Develop-ment and applications of new parenteral adjuvants.III. Factors involved in enhancement of antigenicityof influenza virus by hexadecylamine.' Fed. Proc., 19,219.

JOHNSON, P., NEAL, R. A. and GALL, D. (1963).'Protective effect of killed trypanosome vaccines withincorporated adjuvants.' Nature (Lond.), 200, 83.

LINGGOOD, F. V., STEVENS, M. F., FULTHORPE, A. J.,WOIWOD, A. J. and POPE, C. G. (1963). 'Thetoxoiding of purified diphtheria toxin.' Brit. J7. exp.Path., 44, 177.

MUIR, A. R. (1962). 'An electron microscopic demon-stration ofa surface pattern on the plasma membraneof sectioned intestinal epithelium after saponintreatment.' Nature (Lond.), 195, 1023.

NOLL, H. and YOUNGER, J. S. (1959). 'Virus-lipidinteractions. II. The mechanism of adsorption oflipophilic viruses to water-insoluble polar lipids.'Virology, 8, 319.

PHARMACEUTICAL SOCIETY OF GREAT BRITAIN (1960).The Pharmaceutical Pocket Book, 17th edn, p. 29.Pharmaceutical Press, London.

RAMON, G., RICHOU, R. and STAUB, A. (1937).'L'immunisation au moyen de microbes vivants-bacille diphterique, staphylocoque, bacteridie char-bonneuse-additionnes de substances adjuvantes.'Rev. Immunol. (Paris), 3, 389.

RICHOU, R., JENSEN, R. and BELIN, C. (1964). 'Re-cherches sur la saponine, substance adjuvante etstimulante de l'immunit6.' Rev. Immunol. (Paris), 28,49.

RIVENSON, S. (1958). 'Determination of the adjuvantcapacity of aluminium hydroxide, carboxymethyl-cellulose and saponin in the production of foot-and-mouth disease complement fixing antibodies inguinea pigs.' Gac. vet. (B. Aires), 20, 209; (1959)Biol. Abstr., 33, 3868.

SBARRA, A. J., SHIRLEY, W. and BARDAWIL, W. A.(1962). 'Piggy-back phagocytosis'. Nature (Lond.),194, 255.

WEISS, D. W. and DUBOS, R. J. (1956). 'Antituber-culous immunity induced by methanol extracts oftubercle bacilli-its enhancement by adjuvants.'J. exp. Med., 103, 73.

WOODHOUR, A. F., JENSEN, K. E. and WARREN, J.(1960). 'Antibody response to influenza vaccinescombined with hexadecylamine.' Proc. Soc. exp. Biol.(N.T.), 103, 200.

YOUNGER,J. S. and AXELROD, V. (1964). 'Antigenicityof lipid adsorbed diphtheria and tetanus toxoids.'J. Immunol., 92, 879.

The Adjuvant Activity of Aliphatic Nitrogenous Bases - NCBI

Documents