Eur. J. Immunol. 1977.

7 195-203 Nonspecific immune response genes 195

Maria Siqueira’, Maria B. Esteves’, Nonspecific genetic regulation of antibody responsiveness

Olga M. Ibanez’, Vera C.A. Ferreira,
O.A. Sant’Anna’, Moema H. Reis’ and
G. Biozzi’
DWaflamnto de ImmunolWia, lndtuto
Biologico, Sao Paulo’ and U125 INSERM-
ER 70 CNRS, lnttitut du Radium, Pariso
in the mouse”
Four lines of mice were produced by selective breeding for quantitative
agglutinin responsiveness to flagellar (f) or somatic (s) antigens (Ags) of
Salmonellae : high (H) or low (L) responder lines to fAg and H and L
responder lines t o sAg.
The Salmonellae contained both f and sAgs, the Ag used t o perform the
selection was the Selection Ag and the other was the Associated Ag. The
selective breeding produced a progressive interline separation with an equi-
valent effect for both Ags. After 15 generations (F15) the level of agglutinin
response was about 60 times higher in H than in L responders. About 50 %
of the phenotypic variation of the character investigated is determined by a
group of immune response genes, the rest is due t o environmental factors. The
nonspecific effect of this group of immune response genes was investigated by
measuring the responses to three independent antigens: Sheep erythrocytes
(SE), dinitrophenylconjugated human I& (DNP-HGG) and bovine IgG (BGG).
The selection for fAg response produced an equivalent modification in the
responsiveness to the Associated Ag (97 %) and to BGG (1 30 %). This non-
specific effect was smaller for responsiveness t o SE and DNP-HGG, 58 % and
41 % of the Selection Ag response, respectively.
The selection for sAg response produced a nonspecific modification of re-
sponsiveness of 94 % for the Associated Ag of 74 % for BGG and 63 % for
DNP-HGG. An important exception concerned SE t o which an equal antibody
response is produced in high and low lines of sAg selection.

1. Introduction The phenotypic character investigated “quantitative agglu-

In a preceding study the production of high (H) and low (L)
responder lines of mice by a two-way selective breeding for
quantitative agglutinin responsiveness to flagellar (f) or so-
matic (s) antigens (Ags) of Salmonellae was described.
The non-cross-reacting Salmonella typhimurium (Salm. t m )
and Salmonella oranienburg (Salm. or) were alternated for
immunization of successive generations in order to avoid
the interference of maternally transmitted antibody. Starting
from a population of 75 outbred albino mice, two indepen-
dent selections were carried out. One for agglutinin respon-
siveness to fAg and the other for agglutinin responsiveness
to sAg. Four lines of mice were therefore produced: H and L
responders t o fAg (H/f line and L/f line, respectively) and H
and Lresponders to sAg (H/s line and L/s line, respectively).
tinin response” to f and t o sAgs is submitted t o polygenic
regulation. The heritability realized during the selective
breeding is 0.18 for f and sAgs.
Each Salmonella contained both f and sAgs, the serum agglu-
tinin titers were established independently for each Ag. The
Ag used to perform the selection was called Selection Ag
and the other Associated Ag. The selective breeding for re-
sponsiveness to Selection Ag produced an equivalent modi-
fication in the responsiveness to the Associated Ag. F and
s are non-cross-reacting Ags. Therefore, f and s responses
may be considered as independent characters. The correla-
tion between the response t o f and s Ags is due for 92 % to
additive genetic factors. This strict correlation underlines
the importance of nonspecific factors in the genetic regula-
tion of antibody synthesis to complex immunogens such as
f and sAgs of Salmonellae [ 1 1.

In the present article, we describe the antibody response of

[I 16071
* M. Siqueira, M.B. Esteves and O.A. Sant’Anna were supported in
part by the Conselho Nacional de Desenvolvimento Cientifico e
Tecnologico, Brazil.
Correspondence: Cuido Biozzi, Fondation Curie - Institut du Radium,
Section de Biologie, 26, rue d’Ulm, F-75231 Paris Ctdex 05, France
Abbmirtions: Ag: Antigen f: Flagellar s: Somatic H:High
responders L: Low responders Salm.lm.: Salmonella typhimurium
Sahn.or.: Salmonella oranienburg SE: Sheep erythrocytes DNP-
HCC: Dinitrophenylconjugatedhuman I& BCG: Bovine IgG
iv.: Intravenous i.p.: Intraperitoneal
the four lines of mice: H/f, L/f, H/s and L/s to the three fol-
lowing Ags: Sheep erythrocytes (SE),dinitrophenylated
human IgG (DNP-HGG) and bovine IgG (BGG). The antibody
response to these Ags was measured at various stages of the
selective breeding comprised between the 4th and the 15th
generations (F4 -F15). The modification of antibody re-
sponsiveness to these Ags was compared with that observed
for the Selection Ag. To make this comparison possible, all
the Ags used were administered at optimum immunizing
doses and the antibody responses were measured at the level
of maximum plateau. This steady level of serum antibody
concentration corresponds to a defined period of the immune
response dynamics when the rate of antibody synthesis is
equal t o that of antibody catabolism.
196 M. Siqueira, M.B. Esteves. 0 . M Ibanez et a t
2. Materials and methods
2.1. Mice
The four lines of mice: H/f, L/f, H/s and L/s have been pro-
duced by a two-way selective breeding for quantitative agglu-
tinin responsiveness to f or s Ags of Salmonellae as previously
described [ I ] .
The antibody response t o SE, DNP-HGG and BGG was mea-
sured in populations of mice genetically equivalent t o the F4,
F,, Fa, F,,, Flz, FI4and F,, of the selective breeding. These
genetically equivalent populations have been produced by a
second mating of the same parents chosen for the selective
breeding. The generations tested with the various Ags are
indicated in each experiment. Each Ag was tested in groups
of mice composed of an equivalent number of males and
females and including approximately the same proportion
of individuals from the various families. The response to SE,
DNP-HGG and BCG was measured in groups of 2-month-old
mice. In order t o reduce the impact of environmental factors
o n the interline difference of immune responsiveness, t h e im-
munization and the antibody test of H and L line of the same
generation was made at the same time using the same Ag pre-
paration.
The response t o the Selection Ag and t o the Associated Ag
measured in mice of selective breeding was compared with
the response to the other three Ags in the corresponding ge-
nerations.
2.2. Antigens and immunization
2.2.1. Selection Ag and Associated Ag of Salmonellae
Two intraperitoneal (i.p.) injections of 3.3 x 1 O8 of Salm. tm.
o r 1 x l o 9 of Sulm. or. were given 8 days apart, the anti-f and
anti-s serum agglutinin titer was established 1 0 days after the
second injection. As previously described, this procedure mea-
sures the maximum level of agglutinin response induced by
optimal immunization for both f and sAgs [ 11. The optimal
immunizing doses of the other three Ags used in this study:
SE, DNP-HGG and BGG were determined in preliminary ex-
periments. The maximum antibody response was expressed
as the mean of 2-5 measurements made during the maximum
.
p la t eau
Washed SE (5 x 1 08)suspended in 0.2 ml of saline were in-
jected intravenously (i.v.). Four measurements of serum ag-
glutinins were made during the maximum plateau of t h e re-
sponse comprised between 4 and 20 days after immunization
(see Fig. 4).
HGG was prepared by precipitation of normal human serum
with 18 % sodium sulfate followed by DEAE-cellulose chro-
matography eluted with 0.05 M phosphate buffer, pH 8. The
DNP conjugate was prepared as follows: 300 mg of HGG, dis-
solved in 3 0 ml of 5 %sodium carbonate, was added slowsly un-
der constant stirring t o 0.2 ml of 1,5-difluorc-2,4-dinitrobenzene
diluted 1 /I0in dioxane ( 1 3 0 mglml). After exhaustive dia-
lysis the protein concentration was determined by the micro
Kjeldahl method. The number of DNP groups fixed was de-
termined at 360 nm. The preparations used contained 25-30
DNP groups/molecule.
Eur. I. ImmunoL 1977. 7: 195-203
One mg of DNP-HGG emulsified in 0.2 ml of Freund's com-
plete adjuvant was injected i.p. The maximum anti-DNP re-
sponse was measured 8 and 1 2 days after immunization by
passive hemagglutination of SE sensitized with DNP.
BGG was prepared by precipitation of normal bovine serum
with 1 8 % sodium sulfate followed by DEAE-cellulose chro-
matography, elution with 0.05 M phosphate buffer, pH 8. Fifty
mg of BGG (dissolved in 5 ml of saline) was precipitated with
1.7 ml of 1 % potassium alum, after neutralization with 0.1 N
sodium hydroxide the volume was adjusted t o 10 ml with saline.
Mice were immunized with 3 injections of I mg each of alum-
precipitated BGG. The first injection wasgiven i.v. and the t w o
others i.p. 4 and 8 days later. The maximum antibody re-
sponse was measured by passive hemagglutination 10 days
after the last Ag injection.
2.2.2. Determination of serum antibody titer
The antibody response was measured by direct agglutination
for f a n d sAgs of Salmonellae as well as for SE and by passive
hemagglutination for DNP-HGG or BGG using SE coupled
with these Ags.
The agglutination titer was established using doubling serum
dilutions. It was expressed either as the highest serum dilu-
tion giving positive agglutination or as the log, of this serum
dilution. Individual blood samples were collected with a fine
glass pipette from the retro-orbital venous plexus. After clot-
ting, the serum was separated by centrifugation and kept
frozen at - 20 'C. The individual sera of H and L lines of
each experiment were tested a t the same time using the same
Ag preparation.
Salmonella agglutinin. The suspensions of &lm. tm. and
Salm. or. for the f and s agglutination were prepared by
treatment with formol or ethanol according t o the classical
method [2]. F agglutination: 0.1 ml of doubling serum dilu-
tion in saline was added t o 0.1 ml of a suspension containing
1.4 x 1O9 Salmonella/ml. The agglutination was made in glass
tubes and scored after 2 h at 37 'C. S agglutination: 0.1 m l
of doubling serum dilution was added to 0.1 ml of a suspen-
sion containing 0.9 x 1 O9 Salmonellae/ml. The agglutination
was made in glass tubes and scored after 24 h at 37 'C.
SE agglutinin. The hemagglutination was made in microplates:
0.025 ml of doubling serum dilution in buffered saline con-
taining 1 % of sheep serum was mixed with 0.025 ml of
washed SE suspension containing 1 O8 SE/ml. The agglutina-
tion was scored after 2 4 h at room temperature.
DNPagglutinin. The passive hemagglutination of DNP-sen-
sitized SE was made in microplates according t o the same
technique used for SE. The coupling of DNP t o S E was per-
formed according t o t h e technique used by Del Guercio and
Zola [3].
BCG agglutinin. Glutardialdehyde-treated SE were coupled
with BGG according to the method of Avrameas et al. [4].
The passive hemagglutination was made in microplates with
the same technique used for SE.
2.2.3. Statistics
-
x = mean; u = standard deviation; uz = variance. Standard
error (S.E.) of t h e mean = u Z / awhere N is t h e number of
mice tested.
Eur. J. lmmunol.1977. 7 195-203 Nonspecific immune response genes 197

Significance was assessed by the Student's t-test. The best Selection for responsiveness to Flagellor @(f)
fit of correlations represented in Figs. 3, 5 and 6 was deter-
mined from a least-square linear regression. The degree of
8
.-.
...
H/f
L/f
association between the two variables is measured by the
correlation coefficient.

3. Results

3.1. Antibody responsiveneas to Selection Ag and Associated Ag

The effect of 15 consecutive generations of selective breeding

for responsiveness to f o r sAg of Salmonellae is shown in Figs.
1 and 2. For each selection the mean of the response t o Se-
lection Ag and Associated Ag is shown separately. In the upper
part of the Figures the phenotypic variance of individual re-
sponses t o fAg (Fig. 1) and sAg (Fig. 2) are represented, mea-
sured in each of the four lines resulting from the two selections.
For graphical simplicity, only the variance of the response t o
Salm.tm. is represented in the Figures. The range of the vari-
ance of Salm. or. responses is similar. Both of them show the
. __ 1/40 Llf

34
same trend t o decrease as the selection proceeds.
UImJ Generations
The selective breeding produces a progressive interline di-
vergence in the responsiveness t o both selection Ags f and s. Figure 1. Mean log2 values of agglutinin titers to Selection Ag .f and
This suggests a polygenic regulation of the character inves-
tigated. After 15 generations (F15) the mean agglutinin level
-
to Associated Ag sin the 15 consecutivegenerations of selective breed-
ing for Hand L responsiveness to flagellar Ag of Salmonellae. The in-
is 5 1 and 60-fold higher in H lines H/f and H/s than in L lines dividual variance of the generations tested with Salm. trn. is indicated
L/f and L/s (Figs. 1 and 2). The evaluation of interline diver- in the upper part of the figure (direct agglutinin titers are indicated
for Fo and F15).
gence for the selection Ags is indicated for F4, F6, F8, Flo and
F14 in comparison with responsiveness t o the other Ags (SE,
DNP-HGG and BGG) measured in corresponding duplicated
generations. The results of Figs. 1 and 2 demonstrate that the
interline divergence in responsiveness t o the Selection Ag is Selection for responsiveness to Somatic Agk)
accompanied by an equivalent effect in the response t o the .-.HA

Associated Ag: 48 times in f selection and 52 times in s se-

lection. Since f and s are non-cross-reacting Ags, the effect
of the selective breeding is essentially nonspecific.

In both selections, the progressive interline divergence of the > 3 6 9 12 15

mean agglutinin titer is accompanied by a concomitant re-
duction of the phenotypic variance of the character. This
1 2 ~Reroo;-e

10
to Selection Ag r
,.\,- I/;lOO H/s

means that the lines become genetically more homogeneous

as the selection proceeds. The value of the phenotypic vari-
ance in F15of both f and s selections is about half that found
in the foundation population. Therefore, approximately 50 76
of the phenotypic variance of the foundation population are
genetically determined, the rest being due t o environmental 2 1
effects. 3 6 9 12 15
Response to Asrocioted Aq f ,

The impact of environmental factors on the level of responsive-

ness of each line is responsible for the erratic variations ob- 52 limes
served in successive generations of the two selections repre-
sented in Figs. 1 and 2. Differences in the various batches of
Ag preparations used for immunization or agglutinin test,
and seasonal variations in immune responsiveness may be 5
3 6 9 12 15
1
considered as partially responsible for these environmental m Generations
effects. The measures of agglutinin response in corresponding
generations of H and L lines were always made at the same Figure 2. Mean log2 values of agglutinin titer to Selection Ag s and
associated Ag fin the 15 consecutive generations of selective breeding
-
time using the same Ag batch; therefore, these environmental
factors would have affected both lines in the same way. This for H or L responsiveness to sAg of Salmonellae. The individual
variance of the generations tested with Salm trn is indicated in the
explains the occurrence of parallel erratic variations in the upper part of the fgure (direct agglutinin titers are indicated for Fo
level of response in each line of both selections (Figs. 1 and 2). and F15).
198 M. Siqueira,M.B.Esteves, OM.I h e z et aL
Thus, the best estimation of the genetic effect of selection
is the interline difference between the mean agglutinin titer
of H and L of the same generation. This measure minimizes
the effect of environmental factors.
In Fig. 3, the difference between the mean titer of agglutinin
of H and L is plotted in ordinates against the corresponding
generations in abscissae, for both selections. The best fit
(slope) of the linear regression and the correlation coefficient
are indicated for each selection.
Selection for reqmsweness to flapelbr A g ( t )
! we measured the response t o SE, DNP-HGG and BGG at
Gnuations
Sebction for respon6iuanez.s lo Somatic Ag (s)
r--Linari&onl
WK(i0n A m c i d l d
3 6 9 12 IS
rrmarJl Generations
Figure 3. Response to selection measured by the difference in mean
agglutinin titer between H and L line off and s selection for both
Selection Ag and Associated Ag.
It is evident from Fig. 3 that the rate of response t o selection
is fairly constant during the 15 generations of selective breed-
ing for both f and sAgs. The deviation of the values of each
generation from the regression line is relatively small, as in-
dicated by the high value of the correlation coefficient. The
response to selection for f and sAgs occurs at the same rate
(slope 0.39 and 0.37, respectively). The most interesting
result is that in both selections the slopes of the linear re-
gression for the Selection Ag and Associated Ag are very
close. This means that selective breeding for responsiveness
to one of the two Ags produces an equivalent modification
of responsiveness t o the other non-cross-reacting Ag through-
out the course of selective breeding. The cumulated response
to selection, as represented in Fig. 3, results from the additive
effect of genetic factors nonspecifically affecting the mecha-
nism of antibody synthesis. These genetic factors produce
an equivalent effect on responsiveness t o the two unrelated
Ags.
Ew. J. IIIUIIU~OL1977. 7 195-203
3.2. Antibody responsiveness to other Ags
3.2.1. Introductory remarks
The phenotypic character considered in this study is the
maximum plateau level of serum agglutinins produced by an
optimal immunization. The difference in antibody level
between H and L lines, measured under these conditions,
is an acceptable evaluation of the effects produced by genetic
factors on the immune response t o the Ags investigated.
In order t o extend the study of the nonspecific modifications
of antibody responsiveness produced by the two selections,
different stages of selective breeding for f o r s response. These
three immunogens were injected at optimal doses according
to the immunization procedure used. The antibody responses
were measured during the maximum plateau of agglutinin
levels, as exemplified for SE response in Fig. 4.
3.2.2. SE
The results of Fig. 4 illustrate the dynamics of agglutinin re-
sponse measured in F4 and F,, mice of both f and s selections.
It may be accepted that the maximum agglutinin level re-
mains steady during the period comprised between the 5th
and the 20th post-immunization day. The measure of the
phenotypic character used in the present study is the mean
value of all the antibody titrations made during this period.
Fig. 4 shows that H and L responder mice of f selection are
also separated for agglutinin response t o SE. Contrary t o this,
the same level of SE agglutinin is produced by H and L re-
sponders of s selection. Fig. 4 illustrates how the phenotypic
character has been measured. The complete results of the
agglutinin responses t o SE measured in 6 generations of f
or s selections are compiled in Tables 1 and 2.
The results of Table 1 show that in the 6 generations tested,
H/f mice produced a stronger response t o SE than the cor-
responding generations of the L/f line. The interline difference
increases as the selection proceeds, and it is highly significant
except in the F4 generation. It may be concluded that the
selective breeding for agglutinin responsiveness to fAg of
Salmonellae is also effective in modifying the antibody syn-
thesis t o SE.
In the H/f line the rise of responsiveness t o the Selection Ag
f produced by the selective breeding (Fig. 1 ) is also roughly
reflected in the modification of responsiveness t o SE in the
successive generations shown in Table 1. This parallelism is
not observed in the L/f line which might be due t o the erratic
variations in the level of immune responsiveness produced
in each line by environmental factors. As previously discussed,
the impact of these environmental factors may be reduced by
measuring the interline difference of corresponding generations
This also applies t o some of the results reported in Tables 3-6
concerning the responses of H and L lines t o DNP-HGG and
BGG.
The results of Table 2 demonstrate that the selection for
responsiveness t o sAg does not affect in the same way the
agglutinin response to SE. No significant interline difference
is found in F4, F12 and F14 generations. Moreover, in the other
generations, a small superiority of the L line is observed.
Eur. J. ImmunoL 1977. 7 195-203 Nonspecific immune response genes 199

Risponie lo SE d w( ond L A linai

81 m The antibody response of mice immunized with DNP-HGG

was measured by passive agglutination of DNP-SE. In con-
y sequence, only the antibodies directed against the DNP hapten
were detected, The anti-DNP responses measured in five gene-
- . ~ h 4 6 (I/%) rations for both selections are shown in Tables 3 and 4.
.-.Us ‘66 (11%)

All the generations of mice selected for H responsiveness to

I 1 fAg. (H/f line) show a higher titer of DNP agglutinin than
5 I0 I5 20 5 m 15 20 the corresponding generations selected for L responsiveness
(L/f line). The interline difference in DNP response increases
as the selection proceeds and it is highly significant in all the
generations tested (Table 3). Similar results are obtained in
101
the study of the response t o DNP in the two lines of mice
selected for responsiveness t o sAg. (H/s and L/s lines), as
shown in Table 4. In this case, however, the interline differ-
ence in DNP response is smaller than that found in mice of
fAg selection (Table 3) and becomes highly significant only
after the 8th generation of selective breeding. It may be con-
cluded that both selective breedings for responsiveness to f
and sAgs of Salmonellae are effective in modifying in the same
way, though t o a different extent, the antibody response t o
DNP-HGG.

3.2.4. BGG

Table 1. Agglutinin response to SE of mice selected for H or L responsiveness to fAg of Salmonellae

Generation Hlf line Llf line Aggl units P

No. of No. of SE aggL Ad. H/f :L/f
mice titerr titera) UllitSb)

F4 10 6.2 f 1.45 73 10 5.9 f 1.26 59 1.2 0.6 (ns)

F6 10 6.2 f 1.10 73 9 5.1 f 0.87 34 2.1 0.02
F8 15 8.7 f 1.20 415 16 6.2 f0.80 73 5.7 < 0.001
F 10 16 10.2 f 0.76 1176 16 8.6 f 1.0 387 3.O < 0.001
F 12 14 8.8 f 1.12 445 14 6.82 1.30 111 4.0 < 0.001
F 14 18 8.7 f 1.10 415 18 5.6 f 2.20 48 8.6 < 0.001

a) Logz(Z f U) 5-20 days after immunization.

b) Expressed as the highest reciprocal serum dilution giving positive agglutination.

TaMe 2. Agglutinin response to SE of mice selected for H or L responsiveness to sAg of Salmonellae

Cenaation HI8 line Lls line A@. units P

No. of SE a#. Awl. No. of SE @. Aggl. HIS :L/s
mice titera) unit&) mice titer units
F4 10 6.6 f 1.10 97 10 6.6 f 0.89 97 1.0 n.s.
F6 8 6.2 f 0.76 73 10 7.3 f 0.85 157 0.46 0.01
F8 16 7.3 f 0.80 157 16 8.3 f 1.16 315 0.50 0.01
FlO 16 7.7 f 0.76 207 14 8.4 f 0.90 337 0.61 0.05
F 12 16 7.7 f 0.52 207 15 7.5 f 0.93 180 1.15 n.n
F I4 15 7.3 f 1.43 157 15 7.3 f 0.70 157 1.0 n.s.

a) For explanation see footnotes to Table 1.

200 M. Siqueira, M.B.Esteves, O.M.Ibanez e-t aL Eur. J. ImmunoL 1977. 7: 195-203

Table 3. Response to DNP-HGG of mice selected for H or L responsiveness to fAg of Salmonellae

Generation H/f line L/f tine AggL units P

No. of m. m. No. of Aggl. 4@ H/f :L/f
mice titera) unitsb) mice titera) Unitsb)

F6 10 6.7f 0.72 104 8 5.5 f0.64 45 23 c 0.01

F8 17 7.3 f0.79 157 16 5.6 f 0.88 48 3.3 < 0,001
F 10 16 11.9 f 0.62 3 821 16 10.2 f 1.04 1176 3.2 < 0.001
F 12 14 10.6 f 1.01 1551 12 8.5 f 0.8 362 4.3 < 0.001
F 14 18 12.0 f 0.84 4 0% 18 9.5 f0.9 724 5.7 c 0.001

a) Passive agglutination titer (log2 i f (I) 8-12 days after immunization.

b) Expressed as the highest reciprocal serum dilution giving positive agglutination.

Table 4.Response to DNP-HGC of mice selected for H or L responsiveness to sAg of Salmonellae

Generation H/s line L/s line Aggl units P

No. of A@. 4T$. No. of A@. Ad. Hh :Lls
mice titera) unit&) mice titera) unitsb)

F6 8 6.1 f0.89 68 10 6.1 f0.57 68 1

F8 16 6.3 f 0.85 79 16 5.9 f 0.99 60 1.3 0.25
F 10 15 8.9 f 1.02 477 15 8.3 f 0.92 315 1.5 0.1
F 12 15 9.5 f 1.06 724 16 8.0 f 0.92 256 2.9 < 0.001
F 14 15 9.5 f 1.12 724 1s 7.7 f 1.33 207 3.5 < 0.001

a) Passive agglutination titer (log2 if U) 8-1 2 days after immunization.

b) Expressed as the highest reciprocal serum dilution giving positive agglutination.

Table 5. Response to BGG of mice selected for H or L responsiveness to fAg of Salmonellae

Generation H/f Une L/f line A&lL units P

No. of lregl Md. No.4 Aggl Ira%
mice titera) unitsb) mia titera) unitsb) H/f :Llf
F LO 20 10.3 f 1.03 1260 20 6.2f 2.69 73 17.3 c 0.001
Fn 14 10.0 f 1.30 1024 14 6.3 f 1.43 78 13.1 c 0.001
Fn 18 9.4 f 2.09 675 18 3.8 f 1.73 14 48.2 < 0.001
F1S 16 *
9.7 1.50 831 16 3.4 f 1.42 10 83.1 c 0.001

a) Passive agglutination titer (log2 2f (I) 10 days after immunization.

b) Expressed as the highest reciprocal serum dilution giving positive agglutination.

Table 6 . Response to BGG of mice selected for H or L responsiveness to sAg of Salmonellae

Gemration H/s line Lls line lbgL tmits P

No. of w. No. of &@. &@. HIS: L/s
mice titerr) unit8 mice titerr) UNtS

F 10 19 5.9 f 1.68 59 14 4.3 f 1.20 19 3.1 c 0.01

Fl2 16 10.3 f 1.44 1260 16 9.1 f 1.20 548 2.3 < 0.02
F I4 15 8.8 f 2.68 445 15 6.1 f 2.47 68 6.5 0.01

a) For explanation see footnotes to Table 5.

In all the generations tested, t h e H lines of both selections
presented a stronger response t o BGG than the corresponding
generations of L lines. The difference in BGG responses be-
tween H and L responders is larger in mice selected for anti-f
than for anti-s responsiveness.
As previously explained, it is possible t o reduce the effect of
environmental factors responsible for the erratic fluctuations
of each line b y expressing the results as interline difference
of corresponding generations. We have thus expressed the
results obtained in the study of the three Ags: SE,DNP-HGG
Eur. J. lmmunoll977. 7 195-203 Nonspecific immune response genes 201

TaMe 7. Interline difference in response to SE,DNP-HGG and BGG,of mice selected for H or L responsiveness to fAg of Salmonellae

Gemtion Selection Ag (0 SE DNP - HGG BGG

Agpltita.) %of A d . titera) % of rrepr titera) % of titera) % of
H-L H ~ o Selection H-L line Selection H-L line Selection H-L line Selection
Ag At3 & Ai3

F4 0.8 100 0.3 38 - -

F6 2.2 100 1.1 50 1.2 54 - -
F8 3.6 100 2.5 69 1.7 47 - -
FlO 4.1 100 1.6 39 1.7 41 4.1 100
Fn 5.3 100 2.0 3a 2.1 40 3.7 70
FI4 6.0 100 3.1 52 2.5 42 5 .6 93
F IS 5.5 100 - - - - 6.3 114

a) Agglutinin titer expressed as logz.

Table 8. Interline difference in response to SE, DNP-HGG and BGG of mice selected for H or L responsiveness to sAg of Salmonellae

Generation Selection A8 0) SE DNP - HCC BCG

~ggl.tite& %of w tits& %of ~ g titera)
g ~ %of m.titera) %of
H-L line Sekction H-Lhe Selection H-L line Selection H-L line Selection
Ag & A8 A8

F4 1.5 100 0 0 - - - -
F6 1.7 100 - 1.1 0 0 0 - -
F8 3.2 100 - 1.0 0 0.4 13 - -
F I0 5.2 100 - 0.7 0 0.6 12 1.6 31
FIZ 3.5 100 0.2 5 1.5 43 1.2 34
Fl4 6.0 100 0 0 1.8 30 2.7 45

a) Agglutinin titer expressed as logz.

and BGG in mice of fAg selection (Table 7) and in mice of
sAg selection (Table 8).
In all instances, except for SE response which is not modified
in sAg selection (Table 8), the interline difference in respon-
siveness increases as the selection proceeds from F4 to FI4
generation. Consequently, the percentage of modification
of responsiveness t o the three Ags in relation to that of
selection. The results presented in Fig. 5 demonstrate that
there are good linear correlations between the increase of
interline divergence t o the Selection Ag and that to the other
Ags. The values of the correlation coefficients are high.
I Corr.

Selection Ag tends to remain fairly constant. This percent-

age may be considered as an approximative appreciation of
the nonspecific effect of selective breeding on the response
to each of the other Ags investigated, The comparison be-
tween the data of Tables 7 and 8 indicates that the selection
for fAg responsiveness has a stronger nonspecific effect than
the selection for sAg. This last selection, moreover, does not
operate at all o n the response t o SE.

A more precise comparison of the quantitative correlation

between the responsiveness t o Selection Ag and that t o each
of the other Ags investigated, including the Associated Ag,
is described in Fig. 5 for the fAg selection, and in Fig. 6 for
the sAg selection. The interline difference in the response
to Selection Ag is plotted against interline difference in re-
sponse to each Ag tested, including the Associated Ag. 22+- I
6 I 2 3 4 5 6 7
Selection Ag f interline difference
In order t o increase the accuracy of the comparison, the values IJJmZl ii &@.titer H-L
of interline divergence of the Selection Ag are the theoretical
ones, calculated for each generation on the linear regression
Figure 5. Correlation between the interline differences in response
represented in Fig. 3. This correction reduced the impact of to Selection Ag .f and to the other A g s tested: Associated Ag .s, SE,
environmental factors-responsible for the deviation of a single DNP-HGG and BGG (the slope is calculated from the least-square
generation from the mean value of the cumulated response t o regression line).
202 M.Siqueira, M.B.Esteves, O.M.Ibanez et aL
The slopes of the linear regression measure the nonspecific
effect produced by the selection for fAg o n the responsive-
ness to each of the other unrelated Ags tested. This nonspeci-
fic effect differs according t o the Ag investigated. It is close
t o 100 % for the Associated Ags and BGG. The selective
breeding was therefore as effective in modifying responsive-
ness t o these two Ags as it was for the selection Ag itself.
The nonspecific modification of responsiveness t o SE and
DNP-HGG is 58 % and 41 %, respectively, of that of Selec-
tion Ag . f. The different slopes tend to intercept the axes of
the abscissae at the origin. This suggests that the nonspecific
effect of the selective breeding operates from the beginning
of the interline divergence of responsiveness t o Selection Ag * f.
I s or f, respectively.
2
c
.-
e
2 The progressive separation of H and L lines during 1 5 genera-
0
a tions indicates that the character investigated is submitted t o
0-1’ ,
1
,
2
a,
3
* ,
4
.
5
.6
T
7 ing was measured independently, using three unrelated multi-
Selection Ag s interline difference
m?m X AggLtiter H-L
Figure 6. Correlation between the interline differences in response to
Selection Ag-sand totheother Agstested: AssociatedAg.f,SE,DNP-
HGG and BGG (the slope is calculated from the leastquare regression
line).
The results of Fig. 6 underline the lack of correlation between
responsiveness t o Selection Ag . s and t o SE. In the 6 genera-
tions tested (F4 t o F14), the SE agglutinin levels of the H/s
line are similar or lower than that of the L/s line. This dis-
sociation is limited t o SE since good linear correlations are
found between the response to Selection Ag . s and that to
the other three Ags tested: Associated Ag, DNP-HGG and
BGG. The selective breeding is almost as effective (94 %) in
modifying responsiveness to the Associated Ag. This nonspe-
cific effect measures 63 % for DNP-HGG and 74 % for BGG.
For these two Ags the linear regression intercepts the axes
of abscissae at a value of interline difference for Selection Ag
corresponding to the 6th generation of selective breeding.
This means that either the nonspecific effect of selection is
small and detectable only at a large interline divergence for
the Selection Ag . s, o r that it initiates later at about the 6th
generation of selective breeding. The rather high value of the
slopes is in favor of the last hypothesis.
The results of the study of antibody response to Ags unre-
lated t o the selection Ag may be summarized as follows:
the selective breeding produces a nonspecific effect that de-
pends on both the Selection Ag and each of the other Ags
tested.
Em.J. ImmunoLl977. 7: 195-203
The selective breeding for responsiveness t o fAg of Salmonellae
also modified in the same way but t o a different extent the re-
sponse to the other Ag tested: Associated Ag . s, SE, DNP-HGG
and BCG. The same effects are produced by the selective breed-
ing for responsiveness to sAg of Salmonellae with the important
exception of SE that induces similar antibody response in H
and L lines.
4. Discussion
The results previously published [ 11 and the present data on
Figs. 1, 2 and 3 demonstrate that the selective breeding for
antibody responsiveness t o Selection Ags f or s, produced an
equivalent effect o n the responsiveness t o the Associated Ags
polygenic regulation. Since f and s are non-cross-reacting Ags,
this group of genes regulates the synthesis of antibody of dif-
ferent specificity. These genes are therefore “nonspecific im-
mune response genes”. However, in these experiments the
two independent Ags carried by the same Salmonellae are
simultaneously submitted t o the nonspecific initial steps of
immune response, namely phagocytosis and Ag processing,
operated by macrophages. This particular condition could
contribute t o the total nonspecific effect of the relevant genes
concerning f and sAg responses.
In the present study, the nonspecific effect of selective breed-
determinant Ags: SE, DNP-HGG and BGG in separated popu-
lations of mice genetically equivalent to those comprised be-
tween F4 and F15 of both f and s selections. Under these con-
ditions, it was demonstrated that the fAg selection produced
a nonspecific modification of antibody response t o all the
three Ags investigated (Tables 1 , 3 , 5 , 7 and Fig. 5 ) . How-
ever, in the sAg selection an important exception was found
concerning SE t o which equivalent responses were obtained
in H and L lines while the usual interline separation was ob-
served for the other two Ags: DNP-HGG and BGG (Tables 2,
4 , 6 , 8 and Fig. 6).
The quantitative evaluation of the nonspecific effect of both
f and s selections on responsiveness t o the other Ags has been
expressed by the percentage of correlation between the re-
sponse to each Ag and t o the Selection Ag. Such a compari-
son can be made since all the responses were expressed in
terms of agglutinin titers, measured at their maximum pla-
teau of serum concentration. All the Ags were administered
at optimal dose in relation t o the immunization schedule:
i.v. immunization (SE); i.p. immunization with Freund’s
complete adjuvant (DNP-HGG) and i.p. immunization with
alum-precipitated BGG. It is possible that the extent of the
nonspecific effect of the selective breeding depends in part
on these immunization procedures, namely in relation to the
classes or sub-classes of antibody produced.
A large part of the environmental factors affecting the abso-
lute level of serum agglutinin was eliminated by expressing
the results in terms of interline difference. The reduction of
variance produced by selective breeding indicates that about
50 %of the phenotypic variance in the foundation popula-
Eur. J. Immunol. 1977. 7: 195-203
tion is due t o environmental effects. Supposing that t h e H
and L lines F ~ are
s genetically homogeneous, their environ-
mental variance represents less than 1 5 % of their interline
separation which is due t o additive genetic factors (Figs. 2
and 3). Therefore, the quantitative correlation between t h e
responsiveness t o the Selection Ags and the other Ags investi-
gated, represented in Figs. 5 and 6, is essentially due t o the
effect of nonspecific immune response genes.
The results presented in this article confirm and extend the
demonstration of the existence of nonspecific immune re-
sponse genes by selective breeding experiments for quanti-
tative antibody response t o heterobgous erythrocytes (re-
viewed in [ 6 ] ) .The maximum interline separation was reached
after about 15 consecutive generations of selection. The char-
acter is polygenic and determined by t h e cumulative effect
of about 10 independent loci [5], regulating t h e quantitative
antibody response t o many unrelated complex immunogens
such as: bacteria, viruses, parasite Ags, heterologous proteins,
histocompatibility alloantigens, haptens, etc. [6- 11 1. In spite
of the large number and heterogeneity of the Ags tested, the
nonspecific effect of this group of genes cannot be considered
as a general character since two exceptions have been reported
concerning dextran and levan. I n fact, the H and L lines of
mice are well separated in their response t o bacterial poly-
saccharides such as pneumococcal polysaccharide SIII [ 121
and A or C streptococcal polysaccharides (Eichmann, K. and
Biozzi, G., unpublished results) but give equal responses t o
dextran and levan [ 13, 141.
These two exceptions did not markedly restrict the general
nonspecific effect of the relevant genes, since the t w o poly-
saccharides of repetitive structure are Ags of restricted speci-
ficity submitted t o the control of specific immune response
genes [ 15-1 71, which might affect in some way the effect
o f the nonspecific ones.
In the present study, however, a more important exception
has been demonstrated since H and L lines separated for re-
sponsiveness t o sAg of Salmonellae respond equally t o SE
(Tables 2, 8 and Fig. 6). SE is a complex multideterminant
immunogen submitted t o the control of the nonspecific im-
mune response genes, separated by selective breeding for re-
sponsiveness t o fAg of Salmonellae (Tables 1, 7 and Fig. 5).
The reason for such a n important exception limiting the
general effect of t h e nonspecific immune response genes is
Nonspecific immune response genes 203
presently unknown. However, it opens a new experimental
approach t o the study of the interrelations between specific
and nonspecific immune response genes in the regulation of
antibody synthesis.
Received December 8,1976.
5. References
1 Siqueira, M., Bandieri, A., Reis, US., Sant’Anna, O.A. and Biozzi,
G., Eur. J. ImmunoL 1976.6: 241.
2 Edwards, P.R. and Ewing, W.H., Identification of Enterohcteriacae,
2nd. ed. Burgess Publishing Company, Minneapolis 1964, p. 92.
3 Del Guercio, P. and Zok, E.H., Immunochemistry 1972. 9: 769.
4 Avrameas, S., Taudou, B. and Chuilon, S., Immunochemistry
1969. 6: 67.
5 Feingold. N., Feingold, J., Mouton, D., Bouthillier, Y., Stiffel, C.
and Biozzi, G., Eur. J. ImmunoL 1976.6: 43.
6 Biozzi. G., Stiffel, C., Mouton, D. and Bouthillier, Y., in Benacerraf,
B. (Ed.) Immunogenetics and Immunodeficiency, Medical Technical
Publishing Co., Ltd, Lancaster, England 1975, p. 179.
7 Biozzi, G., Stiffel, C., Mouton, D., Bouthillier, Y. and Decreusefond,
C.,Rogr. ImmunobioL 1971. I: 530.
8 Biozzi, G., Stiffel, C., Mouton, D., Bouthillier, Y. and Decreusefond,
C., J. Exp. Med. 1972.135: 1071.
9 Biozzi, G., Stiffel, C., Mouton, D., Bouthillier, Y. and Decreusefond,
C., Ann. Immuno1.-Inst. Pasteur 1974. 1 2 5 C 107.
10 Liacopoulos-Briot, M.,Bouthillier, Y., Mouton, D., Lambert, F.,
Decreusefond, C., Stiffel, C. and Biozzi, G., Transplantation 1972.
14: 590.
11 Stiffel, C., Mouton, D., Bouthillier. Y., Heumann, A.M.,
Decreusefond, C., Mevel, J.C. and Biozzi, G.. in Brent, L.and
Holborow, J. (Eds.) Progress in Immunology 11, North Holland
Publishing Company, Amsterdam 1974, Vol. 2, p. 203.
12 Howard, J.G., Christie, G.H., Courtenay, B.M. and Biozzi, G.,
Eur. J. Immunol. 1972.2: 269.
13 Howard, J.G., Courtenay, B.M. and Deaymard, C., Eur. J. Immunol.
1974.4: 453.
14 Wiener, E. and Bandieri, A., Eur. J. Immunol. 1974.4: 457.
15 Benacerraf, B. and Katz, D.H., in Benacerraf, B. (Ed.) Immuno-
genetics and Immunodeficiency. Medical Technical Publishing
Co., Ltd., Lancaster, England 1975, p. 117.
16 Blomberg, B., Geckeler, W.R. and Weigert, M.,Science 1972. 1 7 7 :
178.
17 Cramer, M. and Braun, D.G., Eur. J. ImmunoL 1975.5: 823.