The IgA system. I. Studies of the transport and immunochemistry of IgA in the saliva

Abstract

1. Five patients with congenital or acquired agammaglobulinemia, lacking detectable IgA in serum or saliva, were transfused with 1 to 2 liters of normal plasma. In 2 of these patients IgA was demonstrated in parotid saliva collected after transfusion, but in none of the 5 was salivary IgG or IgM found. This observation indicates the selective transport of IgA into saliva.
2. The observation by others of an immunochemical difference between serum and sahvary IgA globulin was confirmed. In contrast to serum IgA, salivary IgA is attached to a protein having antigenicity which migrates as a gamma1 globulin. We have termed this protein component "transport piece".
3. The transport piece has been found in an unbound form in the saliva of persons completely lacking IgA: agammaglobulinemic patients, ataxia-telangiectasia patients, a healthy person lacking IgA, and a newborn infant. Free transport piece still occurs in the normal child's saliva after IgA production begins. By adulthood there is usually no free transport piece in the saliva.
4. Heat-aggregated salivary IgA, like heat-aggregated serum IgA, does not fix complement.
5. Our findings offer support for the view that there is a distinct local antibody system for the protection of the mucous surfaces.

Full text

THE IcA SYSTEM
I.
STUDIES OF T~i~: TRANSPORT AND IMMUNOCHEMISTRY OF IGA
IN Ttil~ SALIVA*
BY MARY ANN SOUTH,$ M.D., MAX D. COOPER,$ M.D., FRANK A. WOLLHEIM,§
M.D., RICHARD HONG, M.D., AND ROBERT A. GOOD,II M.D.
(From the Pediatric Research Laboratories, Variety Club Heart Hospital, and the
Arthritis Unit, Department of Medicine, University of Minnesota, Minneapolis)
(Received for publication 23 November 1965)
The mechanism of resistance to upper respiratory infections is poorly under-
stood. Agammaglobulinemic patients, who are relieved in large part of lobar
pneumonia, septicemia, meningitis, and recurrent urinary tract infections by
gamma globulin, have persistent sinorespiratory tract infections, gastroenteritis,
ileocolitis, and exudative enteropathy. This observation led us to consider the
system involved in immunologic defense of the mucous surfaces in these
patients.
Tomasi and coworkers (1) demonstrated that IgA is the predominant im-
mune globulin in saliva and colostrum despite the fact that IgA is a minority
component of the serum immunoglobulins. Tears (2), gastrointestinal secretions
(2), tracheobronchial washings (3), and nasal secretions (4) have since been
shown to contain IgA as the predominant immunoglobulin. Tomasi (5) has
presented evidence that the IgA in the saliva and colostrum are alike but differ
from serum IgA in that they have a higher sedimentation coefficient, llS
rather than the 7S of serum IgA, and that they have an antigenic determinant
not present in serum.
These observations raise certain important questions. Why is IgA concen-
* Aided by grants from the United States Public Health Service (HE-02085, AI-00798,
AM-07372), the National Foundation, American Heart and Minnesota Heart Association.
Patients were hospitalized and studied under support of United States Public Health Service
grant HE-06314 (Cardiovascular Research Center).
Presented in part at the 49th Annual Meeting of the Federation of American Societies for
Experimental Biology, Atlantic City, April, 1965, and at the Society for Pediatric Research
35th Annual Meeting, Philadelphia, May, 1965.
Special Fellow, National Institute of Allergy and Infectious Diseases, United States
Public Health Service, Department of Pediatrics, University of Minnesota.
§ Instructor, Department of Medicine, University of Minnesota. Present address; Uni-
versity of Lurid, MahnS, Sweden.
[] American Legion Memorial Heart Research Professor of Pediatrics and Microbidogy,
University of Minnesota.
615
on September 22, 2015jem.rupress.orgDownloaded from
Published April 1, 1966

616 IGA SYSamM. I
trated in mucous secretions, particularly as compared with IgG, the major
serum immunoglobulin and a molecule of a size comparable to serum IgA? Is
the
IgA
of secretions a locally produced unique type of
IgA,
basically dif-
ferent from serum IgA; or is there a mechanism for selective transport of a
common IgA molecule into the secretions which is reflected in the immuno-
chemistry of salivary IgA?
It is the purpose of this paper to report findings which shed light on these
questions. We have observed that
IgA
can be selectively transported into saliva
from serum, and that salivary IgA has immunochemical characteristics sug-
gesting that it is composed of two separate immunochemical entities. It appears
that salivary
IgA
differs from serum IgA only in that an additional protein is
attached to the IgA molecule in its passage into the secretions. This additional
protein can occur independently of IgA in agammaglobulinemic saliva and in
saliva of many normal children. Although we have not directly demonstrated
any transport role for this IgA-attached protein, we have found it operationally
useful to call it the "transport piece".
Materials and Methods
Collection oi Samples.--Parofid
fluid (referred to hereafter as
"saUva")
was chosen as the
secretion for study, since it could be easily obtained in large volumes from all our subjects,
and since prior studies of this secretion had been most complete. Collections were made by
means of a Curby cap (6), aided by stimulation with sour candy (Regal-Crown imported sour
fruits, Murray Allen Imports, Inc., New York). Saliva samples were concentrated by negative
pressure dialysis. Serum and saliva samples were tightly capped and kept frozen at --20°F
until analyzed.
Antiserum.--Antiserum
against human IgA was prepared in goats using IgA isolated from
colostrum. In each instance the antiserum was absorbed with serum from a person lacking
IgA. This antiserum gave a single line when reacted in immunoelectrophoresis and in gel dif-
fusion against normal human serum. It gave a reaction of identity against colostrum and
normal saliva, i.e. a single line, with spurring over the serum IgA line (see Fig. 1).
Immunochemical and Physicochemical
Stadie~.--IgA levels were determined by a modifi-
cation of the method of Heremans as previously described (7) using radial diffusion in agar
with incorporated antiserum. IgG and IgM quantitations were performed with the Oudin
tube technique (8). The antisem were prepared in rabbits by immunizing with several IgA
myeloma proteins and with polyclonal IgM, prepared from a pool of rheumatoid arthritis
sera (7). Antiserum against human IgG, prepared in monkeys, was obtained commercially
(Immunology Incorporated, Glen Ellyn, Illinois, lot Cyn 34-1). Pure normal immunoglobulins
prepared from serum as described (7) were used as standards. The standard error with both
methods was -4-10%. Since the rate of diffusion in agar is dependent upon molecular size,
and since the standards were prepared with IgA purified from serum, all the apparent IgA
values in saliva, as compared to serum IgA values, are falsely low, but the comparison among
different saliva samples is valid. The sensitivity of the methods was: for IgA, 1 to 2 rag/100
ml; for IgM, 2 to 5 rag/100 ml; and for IgG, 5 to 10 rag/100 ml.
Double diffusion in gel was done by the method of Ouchterlony (9). Immunoelectrophoresis
was done by the micromethod of Scheidegger (10). Salivary IgA was isolated by the method
of Tomasi (5). The complement-fixing activity of isolated salivary IgA, aggregated by heating
at 63°C for 30 rain, was tested by the method of Mayer (11). 1
This test was kindly performed by Dr. Henry Gewurz.
on September 22, 2015jem.rupress.orgDownloaded from
Published April 1, 1966

SOUTH~ COOPER~ WOLLHEIM~ ItONG~ AND GOOD 617
The IgA in normal saliva was reduced by the method of Fleishman et al. (12). The saliva
was incubated with 2-mercaptoethanol 0.2~ in 0.3 M trihydroxy-methylamino-methane
HC1 buffer, pH 8.2, at room temperature for 1 hr. Iodoacetamide was added in equimolar
amounts and the mixture allowed to stand at 0°C for 1 hr. The saliva was then dialyzed
against sodium borate buffer, pH 8.0.
Infusion Studies.--Six
agammaglobulinemic patients were selected for infusion studies.
They ranged in age from 8 to 41 yr and were of both "acquired" and congenital types. None
had detectable IgA in serum or saliva. A total of 1 to 2 liters of fresh frozen plasma was in-
FIG. 1. Antigenic relationship of serum and salivary IgA, illustrated by immune diffusion
in gel. AS, antiserum against human colostral IgA prepared in goats; wells 1 and 4, serum from
a normal person; well 2, saliva from a normal person; and well 3, IgA isolated from saliva.
All saliva samples in this and subsequent illustrations were parotid secretions and were con-
centrated about X 25. Colostral and salivary IgA contain an antigenic determinant not present
in serum IgA, as illustrated by the spurring of the salivary IgA line over the serum IgA line.
This extra component is referred to in the text as "transport piece".
fused into each of these patients during an interval varying from 1 to 3 days. Specimens of
serum and saliva were collected before the infusions and each day following the last infusion
for a period of 5 days. The serum specimens were analyzed to quantitate the levels of IgA
attained by the infusion, and the saliva was analyzed to discover any IgA, IgG, or IgM trans-
ferred from the infused plasma into the saliva.
RESULTS
When IgA isolated from saliva 2 was tested by Ouchterlony analysis with an
antiserum containing antibodies to colostral IgA, it was partially identical with
serum IgA and it had an extra antigenic determinant as previously demon-
strated by Tomasi (5) (Fig. 1).
Isolated salivary IgA, aggregated by heat, did not fix complement. Similar
positive control experiments carried out with heat-aggregated IgG produced a
2 Saliva refers to parotid secretions throughout these results.
on September 22, 2015jem.rupress.orgDownloaded from
Published April 1, 1966

618
IGA SYSTEM. I
sharp drop in the concentration of whole complement activity in the standard
hemolytic assay.
Studies of Infusion of A gammaglobulinemic Patients with Normal Serum. --
Table I presents the findings of the infusion studies. Only 1 subject (M.L.)
attained a serum level of more than 100 rag% IgA, and IgA appeared in this
subject's saliva at the collections on the 1st, 3rd, and 4th days after infusion.
The serum/saliva ratio was approximately 1000/1. There was no detectable
IgM or IgG by Oudin tube or Ouchterlony techniques in any of these saliva
samples or samples from any of the other transfused patients, so the transport of
IgA was selective. Contamination of the samples with serum could not account
for the presence of IgA since this source would have supplied IgG and IgM as
TABLE I
Infusion of A gammaglobulinemic Patients with Normal Plasma
!
Ag .... glob- A ..... t of Peak ..... Apt ........ AOoP~a.ra~ce
ulinemie of IgA in lg ..
recipient ..... infused level of IgA sahva I ian~a~igdat
cc mg q{,
Congenital sex-linked (M.L.)
Congenital sex linked (T.G.)
Congenital "sporadic" (J.K.)
"Acquired" (L.L.)
Congenital sex-linked (T. L.)
1600
1000
1600
1000
2000
150
75
88
80
90
+++*
+~
0§
0§
0§
* Indicates appearance of 0.1 mg of IgA per 100 ml saliva on 3 different days.
{ The individual specimens collected on 5 successive posttransfusion days were pooled
and concentrated X 225.
§ Each sample of saliva was concentrated 50 to 100 times prior to analysis.
well. Because of difficulty in obtaining cooperation from this child, only small
amounts of saliva could be obtained, and extensive physicochemical analysis of
the secreted IgA could not be carried out. Fig. 2 compares immunoelectropho-
retic patterns of norm al saliva, and the saliva of this agammaglobulinemic patient
before and after infusion. The antiserum used was specific for sermn IgA. The
IgA which appeared in the saliva after infusion produced an immunoelectropho-
retic pattern similar to the normal salivary IgA.
Another agammaglobulinenric patient showed a trace of IgA in the saliva
following infusion of normal serum. Although an IgA line was observed in a
pooled and concentrated posttransfusion specimen, the IgA level was too low to
be accurately measured; still this sample contained no detectable IgG or IgM
and, as in the prior study, supported the concept of selective transport of IgA
into the parotid secretions.
Studies of the Secretion-Specific Gamma A Determinant, "Transport piece".--
on September 22, 2015jem.rupress.orgDownloaded from
Published April 1, 1966

SOUTH, COOPER, WOLLHEIM, HONG~ AND GOOD 619
An untreated agammaglobulinemic patient's saliva, when reacted in agar
diffusion with antibody against purified colostral IgA, gave a distinct precipitin
band. By contrast, no precipitin line formed when the same saliva specimen was
reacted with antibody against purified serum IgA. This precipitin band was
subsequently shown to give a reaction of partial identity with the precipitin
Fio. 2. Immunoelectrophoresis of saliva from a normal person and from an agamma-
globulinemic patient before and after infusion with normal plasma. Wells 1 and 3 are empty;
well 2, saliva from a normal person; wells 4 and 6, saliva of M. L. before plasma infusions;
and well 5, saliva of M. L. after plasma infusion. All the troughs contain serum from rabbits
hyperimmunized with purified IgA myeloma globulin. The cathode is at the left. After in-
travenous infusion of fresh whole plasma (1600 ml) IgA appeared in the saliva collected from
this agammaglobulinemic patient on 3 successive days.
band formed when the antiserum against colostral IgA was reacted with normal
salivary IgA, but complete nonidentity to the reaction produced with normal
serum IgA (Fig. 3). This is the antigenic determinant originally described by
Tomasi et al. (5) 3 as attached to salivary and colostral IgA globulin, and is
referred to as the "transport piece" in this and previous reports (13, 14). This
3 This relationship was confirmed with antiserum kindly sent to us by Dr. Tomasi.
on September 22, 2015jem.rupress.orgDownloaded from
Published April 1, 1966

620 IGA SYSTEM. I
FIG. 3. Antigenic relationship of transport piece to IgA, illustrated by immune diffusion
in gel. AS, antiserum against human colostral IgA prepared in goats; well 1, saliva from a
normal person; well 2, saliva from an agammaglobulinemic person; and well 3, serum from a
normal person. The transport piece, occurring independently in the agammaglobulinemic
saliva, is partially identical to normal salivary IgA but nonidentical to serum IgA. Note also
the character of the normal salivary IgA line. Its curve indicates that it is composed of larger
molecules than is the serum IgA globulin.
FIG. 4. Dissociation of the salivary IgA from the transport piece by mercaptoethanol
reduction, shown by immune diffusion in gel. AS, antiserum against human colostral IgA
prepared in goats; well l, saliva from a normal person; well 2, serum from a normal person;
well 3, saliva from a normal person reduced with 2-mercaptoethanol; and well 4, saliva from
an agammag]obulinemic patient. After incubation with 0A M mercaptoethanol, salivary IgA
is dissociated into two components, one antigenically identical to serum IgA aml the other
antigenically identical to the free transport piece as it occurs in the saliva of agamma-
globulinemic patients.
on September 22, 2015jem.rupress.orgDownloaded from
Published April 1, 1966

SOUTH, COOPER, WOLLHEIM, I-IONG~ AND GOOD 621
substance has subsequently been demonstrated in the saliva of each of 18
agammaglobulinemic patients studied to date, 2 patients with ataxia-telangiec-
tasia who lacked IgA in serum and saliva, a healthy adult with the anomaly
described by Rockey, Hanson, Heremans, and Kunkel (15) who had no IgA in
serum or saliva 4 and a 12-hr-old infant before the first feeding, with no demon-
strable IgA in serum or saliva. In more than 100 determinations we have not
found any saliva sample which lacks the transport piece, although concentrating
the saliva sample is usually necessary for its demonstration.
FIo. 5. Occurrence of free transport piece in the parotid secretions of a child, illustrated
by immune diffusion in gel. AS, antiserum against human co/ostral IgA prepared in goats;
wells 1 and 4, saliva of a child; well 2, saliva of an agammaglobulinemic patient; well 3,
saliva of a normal adult; and well 5, serum of a normal adult. The child's saliva contains free
transport piece, represented by the bands nearest the center well, which is antigenically
identical to the transport piece in the agammaglobulinemic patient's saliva. Newborn babies,
although they have no IgA in saliva, have transport piece.
In the reduced and alkylated form the normal salivary IgA is dissociated to
form two distinct bands when reacted against anticolostral IgA, as had been
demonstrated by Tomasi and associates (5). One of these lines is identical with
the serum IgA line, and the other is identical with the transport piece line in
agammaglobulinemic saliva (Fig. 4). The transport piece has also been found
"free" in normal saliva, most frequently in children's saliva (Fig. 5).
Fig. 6 compares the immunoelectrophoretic patterns of the agammaglobu-
linemic saliva and normal saliva. The free transport piece is shown to migrate
in the gamma1 region. The transport piece determinant is demonstrated in both
normal saliva and agammaglobulinemic saliva, but the IgA determinant in the
normal saliva only.
4 Patient of Dr. Douglas Heiner, Salt Lake City.
on September 22, 2015jem.rupress.orgDownloaded from
Published April 1, 1966

622
IGA SYSTEM. I
FIG. 6. Immunoelectrophoretic properties of free transport piece. Well i, saliva from an
agammaglobulinemic patient; well 2, saliva from a normal person; trough A, antiserum against
human colostral IgA prepared in goats; trough B, the same antiserum absorbed with lyo-
philized normal serum; and trough C, the same antiserum absorbed with lyophilized agamma-
globulinemic saliva. The cathode is at the left. The transport piece, a gammal-migrating
protein, is attached to the IgA in normal saliva but occurs independently in the agamma-
globulinemic patient's saliva. Serum IgA, not illustrated here, has a similar migration.
DISCUSSION
These observations are of interest from several points of view. First, we have
been able to confirm and extend the discoveries of Tomasi and associates which
show clearly that IgA in the saliva and colostrum is immunochemieally dif-
ferent from IgA in the serum. The saliwtry IgA can be dissociated into two
antigenically distinct components by treatment with 2-mercaptoethanol. One of
the components appears immunochemicallv identical to serum IgA. The other
component is immunochemically identical to a protein, the "transport piece",
found in the saliva of patients completely lacking both serum and salivary IgA.
on September 22, 2015jem.rupress.orgDownloaded from
Published April 1, 1966

SOUTII~ COOPER, WOLT.lqq~.1]~, HONO, AND GOOD
623
Thus far the IgA component has been found only in combination with the
transport piece component in the saliva of children and adults. However, the
transport piece is present also in an unbound form in the saliva of most children
and some adults. It may be that transport piece is present in free form in young
children's saliva because of the gradual increase of IgA production during
childhood. The IgA serum levels rise slowly with age (16--18) and reach average
adult levels around puberty (7). It could well be that children have an excess
of transport piece because their formation of the IgA component of the salivary
IgA is relatively deficient when compared to that of older children and adults.
Perhaps of greatest interest in these studies is the finding that patients with
agammaglobulinemia and ataxia-telangiectasia who do not synthesize the serum
IgA nonetheless possess the transport piece in saliva. Thus, from these patients
the transport piece can be isolated and studied without need to consider its
immediate relationship to the serum IgA determinant. Further, agammaglob-
ulinemic patients, laddng a plasma cell system for synthesizing IgA (19) as
well as the other immunoglobulins, present a strong argument for the concept
that the transport piece is formed in cells different from those responsible for
production of the IgA and the other immunoglobulins. This is of particular
interest because the transport piece travels electrophoretically as gammal
globulin.
Immunoh/stochemical studies previously carried out on salivary glands are
consonant with these findings, since Tomasi and associates (5) showed that
the serum IgA component could be identified by the fluorescent antibody tech-
nique in the plasma cells of the salivary gland, whereas the additional com-
ponent immunochemically related to salivary IgA but not to serum IgA was
detectable only in the acinar ceils adjacent to collecting ducts (perhaps the
ductal epithelial cells) of the salivary gland.
In these studies we have further shown that IgA can be specifically trans-
ported into saliva of patients with agammaglobulinemia, an observation
which can be taken as support for the concept that the IgA component in saliva
is the same as that in serum, and is derived also from the plasma ceils. The
transport of IgA from serum to saliva in man is not surprising, since antibodies
administered intravenously had been shown to appear in the secretions of the
mouse (20), rabbit (21), and guinea pig (22), although these antibodies were not
known to be of the Igh type. In our studies it is particularly important to
emphasize that the transport of IgA into saliva occurred in the absence of trans-
port of the other two immunoglobulins, IgG and IgM, even though the latter
two components were present in the serum in higher concentrations after
plasma infusions into agammaglobulinenfic patients than was IgA. These
observations suggest the existence of a specific transport mechanism for IgA.
Such a transport mechanism might, indeed, involve the transport piece, with
production of the transport piece by the epithelial cells, reception of the IgA
on September 22, 2015jem.rupress.orgDownloaded from
Published April 1, 1966

624 IoA $¥STEH. I
from the plasma ceils by specific combination with the transport piece, and
subsequent secretion of the combined molecule into the lumen of the salivary
duct.
The infusion studies in the agammaglobulinemic patients stress another
important point, namely that high concentrations of IgA are necessary to
achieve such transport. Indeed, in our experiments transport of IgA from serum
to saliva was observed in only 2 of 5 patients: in pooled posttransfusion samples
of 1 child, but in 3 of 5 posttransfusion samples in the patient with the highest
serum IgA level. In keeping with these observations, other investigators
(5, 23) failed to demonstrate this transportation after infusion of relatively
small amounts of Ira-labeled IgA. Is the need for high concentrations at the
epithelial cell surface the reason for such large numbers of plasma cells in the
lamina propria of the gut, and for their presence in the interstitial tissue of the
salivary glands and mammary glands? Local production of IgA might be
expected to deliver high concentrations of this globulin directly to the epithelial
cells for subsequent combination with transport piece and delivery into the
secretions. In this regard, it is of interest that Crabb~ (24) has shown that an
average of 80 % of the plasma cells of the intestinal lamina propria are IgA-
producing cells, a much higher proportion of IgA producers than is seen in
lymph nodes and spleen (25, 26). ttochwald et al. (27) showed that C14-1abeled
amino acids are incorporated directly into IgA in salivary gland as well as
mammary gland tissue cultures. Evidence of local specific antibody production
in many species has been reviewed by Pierce (28). This evidence, including
the observations of antibody appearing earlier in secretions than in serum, or in
higher titer, or occasionally exclusively in secretions, speaks strongly for local
antibody production.
Heat aggregated IgA and IgA antibodies from serum do not utilize comple-
ment (29--31). In our study, aggregated salivary IgA also failed to show any
complement-fixing activity. Since saliva and intestinal contents are, by pH and
salt concentration, anticomplementary, a complement-requiring antibody
system would be of no biological advantage in these fluids. The IgA group of
antibodies function either alone or with the aid of biologically active sub-
stances other than complement. The transport piece might furnish such aid, or
might serve either to stabilize the IgA molecule or in some other way make it
better able to act in the milieu of the mucous surfaces.
The experiments of nature represented by patients lacking IgA: the agamma-
globulinemics treated only with IgG, and ataxia-telangiectasia patients, sup-
port the concept that this system has biologic significance. The rare healthy
persons with no IgA in serum or secretions seem to have been able to compen-
sate by delivery of both IgG and IgM into the secretions, as reported by Rockey
and associates (15) and Tomasi and associates (5), and as found in 1 person by
on September 22, 2015jem.rupress.orgDownloaded from
Published April 1, 1966

SOUTH, COOPER, WOLLHEIM, HONG, AND GOOD
625
us. 5 These persons may present a key to greater understanding of the whole
local protective process.
SUM'MARY
1. Five patients with congenital or acquired agammaglobulinemia, lacking
detectable IgA in serum or saliva, were transfused with 1 to 2 liters of normal
plasma. In 2 of these patients IgA was demonstrated in parotid saliva collected
after transfusion, but in none of the 5 was salivary IgG or IgM found. This
observation indicates the selective transport of IgA into saliva.
2. The observation by others of an immunochemical difference between
serum and salivary IgA globulin was confirmed. In contrast to serum IgA,
salivary IgA is attached to a protein having antigenicity which migrates as a
gamma1 globulin. We have termed this protein component "transport piece".
3. The transport piece has been found in an unbound form in the saliva of
persons completely lacking IgA: agammaglobulinemic patients, ataxia-telangi-
ectasia patients, a healthy person lacking IgA, and a newborn infant. Free
transport piece still occurs in the normal child's saliva after IgA production
begins. By adulthood there is usually no free transport piece in the saliva.
4. Heat-aggregated salivary IgA, like heat-aggregated serum IgA, does not fix
complement.
5. Our findings offer support for the view that there is a distinct local antibody
system for the protection of the mucous surfaces.
We wish especially to thank Dr. Bernard Pollara for advice and
encouragement, and to
thank Mrs. Jan Larsen for valuable technical assistance, and Miss Ann Gabrielsen for aid
in
preparing the manuscript.
BIBLIOGRAPHY
1. Tomasi, T. B., Jr., and Zigelhaum, S., The selective occurrence of gamma,A
globulins in certain body fluids,
J. Clin. Invest.,
1963, 42, 1552.
2. Chodirker, W. B., and Tomasi, T. B., Jr., Gamma-globulins: Quantitative rela-
tionships in human serum and nonvascular fluids,
Science,
1953, 142, 1080.
3. Keimowitz, R. I., lmmunoglobulins in normal human tracheobronchial washings,
J. Lab. and Clin. Meal.,
1964, 63, 54.
4. Remington, J. S., Vosti, K. L., Lietze, A., and Zimmerman, A. L., Serum proteins
and antibody activity in human nasal secretions,
J. Clin. Invest.,
1954, 43,
1613.
5. Tomasi, T, B., Jr., Tan, E. M., Solomon, A., and Prendergast, R. A., Character-
istics of an immune system common to certain external secretions,
J. Exp. Meal.,
1965, 121, 101.
6. Curby, W. A., Device for collection of human parotid saliva,
Y. Lab. and Clin.
Med.,
1953, 41, 493.
B Patient of Dr. Douglas Heiner.
on September 22, 2015jem.rupress.orgDownloaded from
Published April 1, 1966

626 IoA SYSTEM. I
7. WoUheim, F. A., and Williams, R. C., Jr., Immunoglobulin studies in six kindreds
of patients with adult hypogammaglobulinemia,
J. Lab. and Clin. Meal.,
1965,
66, 433.
8. Wollheim, F. A., and Williams, R. C., Jr., Studies on the macroglobulins of human
serum. I. Polyclonal immunoglobulin class M (IgM) increase in infectious
mononucleosis,
N. England J. Med.,
1966, 274, 61.
9. Ouchterlony, 0., Diffusion-in-gel methods for immunological analysis,
Progr.
Allergy,
1958, 5, 1.
10. Scheidegger, J. J., Une micro-m6thode d'immuno-61ectrophor~se,
Internat. Arch.
Allergy and Appl. Immunol.,
1955, 79 103.
11. Mayer, M. M., Complement and complement fixation,
in
Kabat and Mayer's
Experimental Immunochemistry (E. A. Kabat, editor), Springfield, Charles
C Thomas, 2rid edition, 1961, 209.
12. Fleishman, J. B., Pain, R. H., and Porter, R. R., Reduction of gamma globulins,
Arch. Biochem. and Biophysics, (Suppl. 1),
1962, 174.
13. South, M. A., Wollheim, F. A., Cooper, M. D., and Good, R. A., Selective trans-
port of IgA and its biologic significance in human saliva,
Fed. Proc.,
1965, 24,
502.
14. South, M. A., Wollheim, F. A., Warwick, W. J., Cooper, M. D., and Good, R. A.,
Local deficiency of immune globulin A in the saliva of patients with chronic
sinopulmonary disease, ].
Pedlar.,
1965, 6"/, 941.
15. Rockey, J. H., Hanson, L. A., Heremans, J. F., and Kunkel, H. G., Beta 2A-
aglobulinemia in two healthy men,
J. Lab. and CUn. Med.,
1964, 63, 205.
16. Roth, N., Sur semiquanfitativen Erfassung der beiden Serum-lmmun-globuline
Beta~A und Beta,M in Neugeborenen- und Kindesalter,
Ann. Paediat.,
1962, 199,
548.
17. Huntley, C. C., and Lyerly, A., Immune globulin determinations in allergic chil-
dren,
Am. J. Dis. Child.,
1963, 106, 545.
18. West, C. D., Hong, R., and Holland, N. H., Iinmunoglobulin levels from the
newborn period to adulthood and in immunoglobulin deficiency states,
J. Clin.
Invest.,
1962, 41, 2059.
19. Good, R. A., Agammaglobulinemia: an experimental study,
Am. J. Dis. Child.,
1954, 88, 625.
20. Fazekas de St. Groth, S., and Graham, D. M., Studies in experimental immunology
of influenza. X. Passive immunity and its enhancement,
Australian J. Exp.
Bid. and Med. So.,
1954, 32, 369.
21. Schubert, O., Diphtheria antitoxin in human and animal saliva,
Travaux de
L'Institut d'Hygi~ne Publique de L'l~tat Tch2coslovaque,
1938, 2, 45.
22. Burrows, W., and Ware, L. L., Studies on immunity to Asiatic cholera. VII.
Prophylactic immunity to experimental enteric cholera.
Y. Infect. Dis.,
1953,
92, 164.
23. Steihm, E. R., Vaerman, J. P., and Fudenberg, H. H., Immune globulin survival
studies following transfusion, American Pediatric Society, 75th Annual Meet-
ing, Philadelphia, May, 1965.
24. Crabb6~ P. A., Carbonara, A. O., and Heremans, J. F., The normal human in-
testinal mucosa as a major source of plasma cells containing gamma A-immuno-
globulin.
Lab. Invest.,
1965, 14, 235.
on September 22, 2015jem.rupress.orgDownloaded from
Published April 1, 1966

SOUTH, COOPER, WOLLHEI~, HONG, AND GOOD
627
25. Mellors, R. C., and Korngold, L., The cellular origin of human immunoglobulins
(72,
71M,
71A); ]. Exp. Med., 1963,
118,
387.
26. Bernier, G. M., and Cebra, J. J., Frequency of distribution of alpha, gamma,
kappa, and lambda polypeptide chains in human lymphoid tissues, J. Imrnunol.,
1965, 95, 246.
27. Hochwald, G. M., Jacobson, E. B., and Thorbecke, G. J., C14-amino acid in-
corporation into transferrin and beta2A-globulin by ectodermal glands in vitro,
Fed. Proc., 1964, 23, 557.
28. Pierce, A. E., Specific antibodies at mucous surfaces, Vet. Rev. and Annotations,
1959, 5, 17.
29. Heremans, J. F., Vaerman, J. P., and Vaerman, C., Studies on the irnmuno-
globulins of human serum. II. A study of the distribution of anti-BruceUa and
anfidiphtheria activities among gamroass, gamrnatM, and gammalA-globulln
fractions, J. Irnrnunol., 1963, 91, 11.
30. Ishizaka, T., Ishizaka, K., Borsos, T., and Rapp, H. J., C'l fixation by human
isoagglutinins: fixation of C'I by gamma G and gamma M but not by gamma A
antibody, submitted for publication.
31. Ishizaka, K., Ishizaka, T., Lee, E. H., and Fudenberg, H., Immtmochemical
properties of human gamma A isohemagglutlnln. I. Comparisons with gamma
G- and gamma M-globulin antibodies, J. Immunol., 1965, 95, 197.
on September 22, 2015jem.rupress.orgDownloaded from
Published April 1, 1966