Journal of Clinical InvestigationVol. 45, No. 5, 1966

The Proteins in Nasal Secretion: A Longitudinal Study ofthe YA-Globulin, yG-Globulin, Albumin, Siderophilin,

and Total Protein Concentrations in NasalWashings from Adult Male Volunteers *

ROGERD. ROSSEN,t ARTHURL. SCHADE,+ WILLIAM T. BUTLER, ANDJULIUS A. KASEL

(From the U. S. Department of Health, Education, and Welfare, Public Health Service, Na-tional Institutes of Health, National Institute of Allergy and Infectious Diseases,

Laboratory of Clinical Investigations and Laboratory of Infectious Diseases,Bethesda, Md.)

The long-standing observations that respiratorysecretions contain virus-neutralizing substances(1-3) have recently been extended by the demon-stration that antibody activity associated with theyA-globulin in nasal secretion may protect againstrespiratory virus infection (4). Since protectionwas most clearly seen in individuals with relativelyhigh titers of nasal antibody, and since yA-globu-lin is known to be selectively accumulated in nasalsecretion (5, 6), it seems possible that local con-centrations of antibody globulin may be determinedby at least two distinct but interrelated processes,1) those concerned with synthesis of specific im-munoglobulin, and 2) those involved in its deliveryinto the nasal secretions. The present investiga-tion explores mechanisms involved in the deliveryof immunoglobulin by examining the relationshipbetween the mucus content and the concentrationsof yA-globulin, -yG-globulin, albumin, siderophilin,and total protein in nasal washings from normalvolunteers.

MethodsClinical and laboratory evaluation of volunteers. Sub-

jects were 15 adult males who participated in an adeno-virus vaccine trial and were not inoculated with liverespiratory viruses. Throat and rectal swab specimensfor virus isolation and nasopharyngeal swab specimens

* Submitted for publication November 22, 1965; ac-cepted January 26, 1966.

An abstract of some of the findings has been published(Clin. Res. 1965, 13, 298).

t Address requests for reprints to Dr. Roger D. Rossen,Clinical Immunology Section, Laboratory of Clinical In-vestigations, National Institute of Allergy and InfectiousDiseases, Bethesda, Md. 20014.

t Laboratory of Infectious Diseases, Bethesda, Md.

for bacterial culture were collected and processed bypreviously described methods (7).

Method of collecting nasal washings. The method ofperforming nasal washes has been described (6). Inbrief, it required instillation of 10 ml of lactated Ringer'ssolution into the nasal passages, homogenization of theeffluent, and separation of the soluble portion from themucus and cellular debris by centrifugation. The vol-umes of the mucoid sediment and the supernatant weremeasured, and the two were stored separately at - 600 C.All studies of nasal wash protein constituents were doneon the supernatants.

Protein determinations. Total protein concentrationwas determined by the biuret method (8). Protein ni-trogen contents of several samples were also tested bydigestion and nesslerization. Total protein values (N X6.25) by this technique were one and, one-half to twotimes as great as those obtained by the biuret method.A single pool of nasal washings from five volunteers,on the other hand, showed identical protein concentrationsby biuret and micro-Kjeldahl methods (N X 6.25).

Concentrations of oyA-globulin, -yG-globulin, and albu-min were determined by the single radial diffusion methodin agar (9). The following conditions were observed.

Antiserum was brought to the appropriate dilution in0.2 M sodium phosphate buffer, pH 8, and incorporatedinto a 1% agar layer 1.5 mmthick. The antigen was al-lowed to diffuse at 4 to 60 C from wells 1.5 mmin diam-eter. At the end of the incubation period, the diameter ofthe leading edge of the precipitin ring surrounding thewell was read to the nearest 0.1 mm. Serial dilutions ofa standard protein solution were included in each agarplate as controls.

For yA-globulin determinations, the agar containedgoat antihuman yA-globulin diluted 1:12.1 This antise-rum was specific for 'yA-globulin in immunoelectrophore-sis against whole human serum at several test dilutions.The yA-globulin used as a standard was prepared frompooled concentrated nasal washings by chromatographyon DEAEcellulose, as described by Tomasi, Tan, Solo-

1 Lot GP4-65, Hyland Laboratories, Los Angeles,Calif.

768

NASAL SECRETIONPROTEINS

mon, and Prendergast (10). When nasal washings forthe pool were selected to contain minimal concentrationsof yG-globulin and no detectable siderophilin, the pro-

tein eluted with 0.1 M sodium phosphate, pH 6.2, gave a

single line characteristic of 'yA-globulin on immunoelec-trophoresis against antiwhole human serum. Trace con-

tamination, however, by a protein not 'yA-globulin, couldbe demonstrated with antisera prepared against wholenasal washings. This preparation of 'yA-globulin was

heterogeneous by analytical ultracentrifugation, its prin-cipal component having an Sm. of 11.1 S at a proteinconcentration of 1.8 mg per ml. The sedimentation prop-

erties of this purified 'yA-globulin were compared withthose of yA-globulin in 14 samples of whole nasal washstudied by density gradient ultracentrifugation and were

found to be quite similar (11). The minimal detectableconcentration of 'yA-globulin was 0.10 mg per ml 0.03(2 SD).

Concentrations of 'yG-globulin were determined in agar

by using goat antihuman -yG-globulin in a final dilution of1: 30.2 At dilutions lower than 1: 4 this antiserumshowed trace activity by immunoelectrophoresis against-yA- and yM-globulins. A pool of yG-globulin, twicechromatographed on DEAE cellulose (12) after precipi-tation of extraneous plasma protein with 2-ethyoxy-6, 9diamino-acridine-lactate (13), was used as the standard.This preparation of 'yG gave a single line by immunoelec-trophoresis against antiwhole human serum. It was ho-mogeneous with a sedimentation coefficient of 6.8 S bysucrose density gradient ultracentrifugation. The mini-mal detectable concentration of yG-globulin was 0.10 mg

per ml 0.08 (2 SD).For albumin determination, the agar contained a 1: 12

dilution of goat antihuman albumin.3 By immunoelectro-phoresis, this antiserum also developed a line for a, gly-coprotein at dilutions less than 1: 4. The albumin stand-ard had a sedimentation coefficient of 4.7 S by densitygradient ultracentrifugation; it was 99.5%o pure by paper

electrophoresis, but on immunoelectrophoresis a tracecontamination with a, glycoprotein and orosomucoid couldbe demonstrated. The minimal detectable concentration ofalbumin was 0.03 mg per ml 0.02 (2 SD).

Siderophilin concentrations were determined by a modi-fication of the radioactive precipitin method of WMood-worth and Schade (14). A series of standard solutionswas made containing different known amounts of un-

labeled siderophilin and a constant amount of 'Fe-labeledsiderophilin. Rabbit antisiderophilin was then added tothese standard solutions until further addition of anti-serum no longer reduced the radioactivity of the super-

natant of this antigen-antibody mixture. A standardcurve was constructed that related the quantity of un-

labeled material in the standard solutions to the quantityof antiserum required to precipitate the mixture of la-beled and unlabeled siderophilin. Nasal wash siderophi-lin concentrations were then estimated by measuring thequantity of this rabbit antiserum required to precipitate

2 Lot GP5-65 from Hyland Laboratories.3 Lot GP1064F from Hyland Laboratories.

the radioactivity in a sample of nasal wash to which hadbeen added a known amount of 'Fe-labeled siderophilin.Minimal detectable concentration was 0.003 mg per ml ±0.003 (range).

Measurements of total protein, siderophilin, yA-globu-lin, yG-globulin, and albumin concentrations were per-formed directly on the supernates of all nasal wash speci-mens. In addition, measurements of -yA-globulin, yG-globulin, and albumin were performed after the specimenshad been dialyzed against distilled water, lyophilized, andreconstituted at one-tenth the original volume with nor-mal saline. We found, however, that the corrected con-centrations of -yA-globulin, yG-globulin, and albumin de-termined on the lyophilized portion were, respectively,40%o, 70%o, and 10% less than the concentrations meas-ured directly on the unlyophilized portion of the samespecimens. Because of this, linear plots were con-structed that related the concentration of each of the pro-teins in the lyophilized portion to its concentration in theoriginal specimens. These graphs were used to estimatethe actual protein values in specimens in which no pro-tein was detectable on direct measurement. These cor-rections needed to be applied to only 3 of the 42 speci-mens tested for -yA-globulin and albumin, but to 33 of 42specimens tested for yG-globulin.

Immunoelectrophoresis was performed as describedpreviously (6). All specimens were tested under identicalconditions against the same lot of antiserum,4 so thatthreshold concentrations for detection of the various se-rum proteins were the same for all specimens.

Test for occult blood. A portion of the centrifugedsediment was tested by the guaiac method.

ResultsTotal protein content of nasal wash specimens

Figure 1 shows the geometric means + 95 %o con-fidence limits of the total protein concentrations ofall specimens collected from each of the 15 volun-teers. The total protein concentration of serialsamples from any single volunteer varied over anarrow range, and the mean protein concentrationof specimens from any individual tended to be thesame whether his guaiac-positive or his guaiac-negative samples were used exclusively in makingthis calculation (p < 0.01, rank correlation test)(15). On the other hand, there wvas up to a five-fold difference in the mean total protein concentra-tion from one volunteer to another. Day-to-dayvariation in the protein concentration of the nasalsecretion was evaluated by assigning each speci-men on each of the days of the study a rank scorefrom 1 to 15 based on the total protein concentra-tion of the specimen. Evaluation of these rank

4 IEP lot 3802 UI from Hyland Laboratories.

769

ROSSEN, SCHADE, BUTLER, AND KASEL

VolunteerNo.12

14

3

(7) (4)

(2) 18)

1(4) $(78 . _,

(6) (5)

(I) (4)

13 a*(8)

(5) (6)

4 (4) (7)9 (5) (6)

5 13) (8)

15 (7)

2

10 es

(10)

7 o., .,(7) (4)

6 A. A

(I) (10)

(3)

Geometric meanI I-I All specimens from a

single volunteer 95%confidence limits

A Guaiac positive specimens* Guaiac negative specimens

( ) Number of specimensincluded in mean

0 04 08 1.2 1.6 2.0 2.4TOTAL PROTEIN (mg/ml)

FIG. 1. MEANTOTAL PROTEIN CONCENTRATION(BIURETMETHOD) OF ALL NASAL WASHINGSCOLLECTEDFROMEACH

OF THE VOLUNTEERSOVER THE 27-DAY PERIOD OF OBSERVA-

TION.

scores showed that there was a highly significanttendency for the protein concentrations of samplesfrom any one volunteer to maintain the same rankthroughout the study (Kendall coefficient of con-

cordance, 14 degrees of freedom, chi square =

81.16, p<<0.001) (16). Separate evaluation ofonly guaiac-positive or guaiac-negative specimensshowed the same day-to-day constancy in the rela-tive protein concentration of specimens from a

single volunteer.To evaluate the relationship between the pro-

tein concentration of the nasal wash specimens andthe quantity of secretion collected by the nasalwash procedure, the volume of mucoid sedimentrecovered after centrifugation was used as an in-dex of the amount of nasal secretion present in thespecimens. Each specimen was given a rank score

based on this volume. The scores representingmucus volume were also quite constant from dayto day (Kendall coefficient of concordance, 14 de-grees of freedom, chi square = 43.32, p < 0.001).Moreover, comparison of the rank scores repre-

senting. mucus volumes with those representingprotein concentration indicated that the amount ofprotein in the nasal wash sample was closely re-

lated to the volume of mucoid material recoveredwith the specimen (r = 0.836, p < 0.001, rankcorrelation test on sums of ranks for mucus vol-ume and total protein).

Contribution of plasma to the nasal secretion pro-tein

Immunoelectrophoretic studies. If transudationof plasma were unselective and -a major source ofprotein in nasal secretion, one might expect to

identify a large number of plasma protein com-

ponents most frequently in specimens with hightotal protein concentrations. On the other hand,if minute breaks in capillary integrity were chieflyresponsible for the plasma protein in nasal secre-

M Guoioc negative specimens No. above bars = No. of specimens= Guoiac posit ive specimens

Mean Total Protein 0.24- 0.50 mg/ml2 12 (53 specimens from vol.#2,6,7,10, 15)

Meon Total Protein 0.86- 1.51 mg/ml100 _12e7 258 (54 specimens from vol.# 3,8, 12,13, 14)

22~

Z_ 3 2 26 282 Z Z -

50

FIG. 2. FREQUENCYWITH WHICH 12 PLASMA PROTEIN

COMPONENTSWEREIDENTIFIED BY IMMUNOELECTROPHORESIS

IN NASAL WASHINGSFROMVOLUNTEERSWITH LOW (TOP

PANEL), MODERATE(MIDDLE PANEL), AND HIGH (BOTTOM

PANEL) NASAL WASH PROTEIN CONCENTRATIONS. The

protein components are listed in order of the frequency

with which they were detected in nasal wash. The col-

umn at the extreme right indicates the frequency of posi-

tive reactions for occult blood among the specimens in

each group. Volunteer 11 was omitted because the n~um-ber of guaiac-negatiVe samples was inadequate.

770

(f)I--

LAJ

cn:zCl

LLI

C>C=a-

-:Im(n.:x-iCl-

LL-C>

C>I--<c.2L-

I--

LLJm

LA-C>

NASAL SECRETIONPROTEINS

tion, one might expect that only guaiac-positivespecimens would frequently contain a great num-

ber of plasma components. Immunoelectropho-retic studies (Figure 2) demonstrated that therewas no significant difference in the frequency withwhich plasma protein components were identifiedin guaiac-positive and guaiac-negative nasal wash-ings (p 0.10, chi square test) (15). Althoughspecimens with high total protein concentrationsfrequently did contain a large number of plasmaproteins, components such as albumin, yA-globulin,and yG-globulin were identified in most specimensirrespective of their total protein concentration.

Quantitative evaluation. To evaluate the aboveobservations further, we determined the albu-min, yG-globulin, yA-globulin, and siderophilinconcentrations of three guaiac-negative specimensfrom each volunteer. Specimens were chosen togive, where possible, a sampling of days duringthe month of study. The mean values for thosedeterminations are shown in Table I.

Of the four proteins tested, yA-globulin was

present in the highest concentrations and accountedfor a considerably greater percentage of the totalprotein (21 to 53%o) than in serum (17). Con-centration of yG-globulin varied from 5 to 14%o,whereas albumin levels were 2 to 9%o of total pro-

tein. Mean siderophilin concentrations extendedfrom 0 to 2.5%o of the total protein. The average

quantity of total protein accounted for by these

four components was 45 %, and it ranged from31%o in specimens from Volunteer 3 to 73%o inthose from Volunteer 5. In all cases, a substantialportion of the protein remained unclassified.

When the mean concentrations in Table I werecompared, total protein correlated closely with theconcentrations of yA-globulin (r = 0.869), yG-globulin (r = 0.776), and albumin (r = 0.662)(p <0.01, rank correlation test), whereas therewas no correlation between any of these and themean siderophilin concentration. Moreover, whenindividual values for yA-globulin or albumin inserial nasal wash specimens from a single volun-teer were examined, they exhibited the same tend-ency to remain constant from day to day as did thetotal protein [analysis of variance (15), F ratiofor yA-globulin = 3.29, p = 0.01; F ratio for al-bumin = 2.31, p < 0.05]. Individual values foryG-globulin and siderophilin in repeated samplesfrom a single volunteer, however, showed greatervariability, and no day-to-day constancy in theirconcentrations was demonstrated (F ratio < 1.50,p > 0.05, analysis of variance). Indeed, therewas no correlation between the siderophilin con-

centration and that of any of the other measuredcomponents in individual specimens (p > 0.05,rank correlation test).

Table II shows a comparison between the albu-min and siderophilin concentrations in specimensfrom each of 14 volunteers. When detected,

TABLE I

Geometric mean -yA-globulin, -yG-globulin, albumin, and siderophilin concentrations inguaiac-negative nasal wash from 14 volunteers*

No. of TPspecimens accounted

Volunteer examined Mean TPt -yA-globulint -yG-globulint Albumint Siderophilint for

lg/ml %12 3 1,752 481 (27) 90 ( 5) 81 (5) 10 (0.6) 388 3 1,595 592 (37) 105.( 7) 50 (3) 9 (0.6) 473 3 1,400 325 (23) 74 ( 5) 27 (2) 5 (0.4) 31

13 3 1,250 277 (22) 58 ( 5) 60 (5) 11 (0.9) 3214 2 895 252 (28) 63 ( 7) 79 (9) 19 (2.1) 4615 3 785 230 (29) 65 ( 8) 34 (4) 9 (1.1) 43

1 3 765 235 (31) 55 (7) 23 (3) 0 (0 ) 416 3 606 176 (29) 42 ( 7) 26 (4) 5 (0.8) 419 4 590 205 (35) 31 ( 5) 24 (4) 15 (2.5) 474 3 578 216 (37) 71 (12) 33 (6) 11 (2.0) 577 3 500 106 (21) 31 ( 6) 19 (4) 9 (2.0) 33

10 3 478 113 (24) 52 (11) 23 (5) 7 (1.5) 412 3 397 212 (53) 25 ( 6) 24 (6) 2 (0.5) 665 3 370 178 (48) 53 (14) 32 (9) 6 (1.6) 73

* Volunteer 11 was omitted because the number of guaiac-negative samples available was inadequate.t TP = total protein.I The number enclosed in parentheses is the per cent of TP; the other number is the mean concentration, expressed

as micrograms per milliliter.

771

ROSSEN, SCHADE, BUTLER, ANDKASEL

TABLE II

Concentration of albumin and siderophilin in individual guaiac-negative nasal wash specimens

Day [Sider- Day [Sider-Volun- of Albu- Sider- ophilin]/ Volun- of Albu- Sider- ophilin]/teer* study min ophilin [Albumin]t teer* study min ophilin [Albuminjt

,ug/mi pg/mi12 1 158 26 0.16 6 8 23 7 0.30

2 33 0 12 37 012 100 18 0.18 14 20 7 0.35

8 8 75 17 0.23 9 1 20 11 0.5512 50 18 0.36 8 40 18 0.4527 33 0 21 18 18 1.00

3 8 28 0 4 1 33 010 14 0 12 40 39 0.9818 50 30 0.60 18 28 18 0.64

13 12 75 0 7 8 22 021 26 18 0.69 12 16 18 1.1327 108 39 0.36 21 20 22 1.10

14 2 220 39 0.18 10 1 18 11 0.6127 28 9 0.32 8 33 18 0.55

10 20 0

15 2 26 18 0.69 2 8 33 04 37 0 12 20 3 0.158 40 22 0.55 14 20 0

1 8 33 0 5 1 58 17 0.2910 16 0 4 20 7 0.3518 24 0 8 28 0

t Ratios in normal serum = 0.08; serum concentrations* Volunteer 11 omitted (see first footnote to Table I).

siderophilin levels often approached or exceededalbumin concentrations. In each instance, sider-ophilin to albumin concentration ratios were higherthan would be expected if passive transudation gov-erned the delivery of these proteins into nasalsecretion equally. Concentrations of -yG-globulincommonly exceeded albumin levels as well (TableI). The mean -yG-globulin to albumin concentra-tion ratio in 43 individual nasal wash specimenswas 1.8 (range, 0.6 to 5.6), whereas a concentra-tion ratio of 12/40 mg per ml (0.3) might beconsidered normal for whole serum (18). Neithersiderophilin nor yG-globulin, however, approachedthe uniformly high concentrations attained byyA-globulin. This protein was the only measuredcomponent whose concentration in nasal wash rel-ative to total protein was consistently higher thanits plasma concentration.

Analysis of clinical data on the volunteers

There was no association between the mean nasalwash protein level and history or present evidence

= albumin, 40 mg per ml; siderophilin, 3 mg per ml.

of sinusitis, evidence of acute upper respiratorydisease, or status of nasal bacterial or viral florain any of the volunteers (Table III).

Discussion

In serial nasal washings from any one volunteer,both total protein concentration and volume of mu-coid material remained constant over an intervalof nearly 1 month, whereas protein levels variedas much as fivefold from one person to another.The observation that specimens from volunteerswith high nasal wash protein concentrations tendedalso to have large volumes of mucoid sedimentsuggests that protein concentration was a functionof the amount of nasal secretion recovered by thecollection procedure. Presupposing that the vol-ume of mucus in the specimen bore a constant re-lationship to the quantity of secretion in the nasalpassages, one may postulate that the consistentdifferences among volunteers in nasal wash pro-tein and mucus content were due to differences inthe amount of nasal secretion available for collec-

772

NASAL SECRETION PROTEINS

o4 C4o o oooo

oc~4 00 0 0 0 0o o oC

_C _' _ e' CU Ce CU - _~ 0 _ _U _U

bl O bC b~ b bb; b; bO ~ ; b4)4 4)4) 4)4)4))4) ) 4) V 4)4)4)

z z z z z z z z z z z z z

oCC)~~~~~~~~~~~~~~~~~~C

0 m '0 0000 = 0 0 0 .0 0 0> 00

_ -7 ~~~~~~0C, CU

*0

cIS E4 co)4) -c)E*4)c o

C0 CCCUC0 C 5 C F00

E E E E E

Cd-d. & I.b X N.e

\ ~ ~ ~ 4U)4_)ooo s ) t e C

00 0 -La 0 0 0 0 0 U0 0 0 00 0

00 0 o 00 0 0V 0000

U)*e00 _ e _ * o U)w )0 Ot-t 0

04 00

773

*v

,za

4.

4)C"

CUt

004

0 0

0 0.0 )

0 U)00

4)

o

.4)

'0

ANa

0.

na0

c °U

'0

,0

Cc >

g

0.

0)~

o..

cdEc

0C0

o

>)1.*

0

E00

140

Cd

0

0

0

Ci

0

a

0

4)

0C

0

0

0

40

H

CU

O

000

V

A U'

Z

I

0

0

C.0

0t

2._

OU X

0. 0

* 4)f a-

D

ROSSEN, SCHADE, BUTLER, ANDKASEL

tion. There was no evidence that collection of se-cretions was consistently more complete in somevolunteers than in others. No volunteer showeda special tendency to swallow the nasal washings,and the technique and conditions for collectionwere the same in all. The nearly equivalent pro-tein concentrations and composition of guaiac-positive and guaiac-negative specimens from anysingle individual suggest, moreover, that smalltraces of whole blood did not contribute to thehigher protein concentrations or larger mucusvolumes of specimens from some volunteers. Itappears likely, therefore, that these findings repre-sent one aspect of constitutional variation, possiblydue to differences among individuals in the produc-tion or retention of secretions in the nasal passages.However, other factors, such as differences in thephysicochemical composition of secretion or innasal anatomy may have contributed to the differ-ences observed.

Of the many plasma protein components foundin nasal washings, yA-globulin, yG-globulin, albu-min, and siderophilin accounted for slightly lessthan one-half of the total protein. Since the im-munoelectrophoretic studies indicated that otherplasma proteins probably do not contribute con-sistently and significantly to nasal secretion pro-tein, a considerable portion remains unclassified.The contribution of proteins not recognized byantisera to whole human serum is currently underinvestigation.

Of the four measured components, yA-globulinaccounted for the greatest proportion of the totalprotein (21 to 53%), and one may question, there-fore, whether an accurate estimation of the con-centration of this protein was obtained. It hasbeen demonstrated that yA-globulin, both in sera(19) and parotid and mammary secretions (10),may exhibit considerable size heterogeneity. Ifthe distribution of -yA-globulin molecules in ourreference standard were different from that of thespecimens tested, then the estimation of yA-globu-lin concentration might not have been valid. How-ever, this possibility seems unlikely, since the ref-erence standard was a preparation of yA-globulinpurified from pooled nasal washings with sedi-mentation properties similar to those of yA-globu-lin in whole nasal wash (11).

The demonstration that yA-globulin in nasalsecretion may be present in concentrations con-

siderably higher in relation to total protein than inserum (17) extends the work of Remington,Vosti, Lietze, and Zimmerman (5) and suggeststhat the immune system of the nasal tissues maybe similar to that of parotid and mammary glands,as described in the recent investigations of Tomasiand associates (10). Their work, confirmed inpart by McFarlin, Strober, Wochner, and Wald-mann (20), suggested that the parotid gland mayparticipate in the synthesis of yA-globulin dis-charged into its secretions. However, transportof this immunoglobulin from plasma with subse-quent modification of the protein by the glandcould not be wholly excluded.

Although a direct relationship between thequantity of immunoglobulin in nasal washings andresistance to infection has not been demonstratedin this study, it has recently been shown elsewhere(4) that protection against rhinovirus infectioncorrelates with the titer of antibody activity re-covered in nasal washings as well as the titer ofantibody in serum. Nasal washings of a few indi-viduals who became infected failed to reflect theirhigh serum antibody levels. The present worksuggests that, quite apart from variations in theability to synthesize specific globulin, differencesamong individuals in the quantity of antibody lo-cally available may be determined by factors thatregulate the production or retention of nasalsecretion.

In contrast to yA-globulin, nasal wash concen-trations of yG-globulin, albumin, and siderophilinrelative to total protein were less than in serum.Whereas the mechanisms by which these are incor-porated into nasal secretion are also not under-stood, it is evident that if they are passively in-corporated from plasma, the barrier to transuda-tion must be less restrictive for both siderophilinand yG-globulin than for albumin. If other fac-tors that influence the passage of proteins acrosscell membranes are equal, the smaller molecularsize of albumin (60,000 mol wt) compared tosiderophilin (90,000 mol wt) or -yG-globulin (150,-000 molwt), and the higher concentration gradi-ent for albumin across the nasal vascular mem-branes would be expected to produce much lowersiderophilin to albumin and .yG-globulin to albu-min concentration ratios than were observed.Since the immunologic method used to measurenasal wash albumin levels when tested against sera

774

NASAL SECRETIONPROTEINS

gave values 25% higher than a standard method(21) utilizing the dye-binding properties of albu-min, it is improbable that the ratios observed re-sulted from any underestimation of nasal washalbumin concentrations.

The failure to detect siderophilin consistentlyin serial specimens from the same individual andthe lack of correlation between the siderophilinconcentration and the concentrations of yA-globu-lin, -yG-globulin, albumin, or total protein in indi-vidual specimens suggest the possibility that tis-sues, such as those of the paranasal sinuses, thatmay discharge secretions intermittently are chieflyresponsible for the incorporation of siderophilininto nasal secretion.

Although a biological role for siderophilin innasal secretion is not defined, it is reasonable toassume that in this environment it serves as a non-specific antibacterial agent. In its iron-free form,siderophilin exerts a broad spectrum of bacterio-static activity that is concentration dependent onlyin that the amount of siderophilin must be suffi-cient to chelate all soluble iron present in itsmilieu (22). Thus, even at the low concentrationsof siderophilin present in nasal secretion, it is pos-sible that this protein inhibits the development ofmany pathogens with restrictive nutritional ironrequirements.

Summary

Eleven nasal wash specimens were collected fromeach of 15 volunteers during a 27-day period. Themucus volume and total protein concentration ofthese specimens were determined, and each speci-men was studied by immunoelectrophoresis tocharacterize qualitatively its content of plasmaproteins. Quantitative estimations of -yA-globu-lin, 'yG-globulin, albumin, and siderophilin concen-trations were made on several specimens from eachof the volunteers.

The total protein concentration and mucus vol-ume of serial nasal wash specimens from a singlevolunteer were relatively constant over the testperiod, whereas there was considerable variation inprotein concentration among samples from differ-ent volunteers. Differences in protein levelsamong volunteers were not correlated with acuteor chronic respiratory illness or marked differencesin nasal bacterial flora.

Between 21 and 53%o of the total nasal wash pro-tein was accounted for by yA-globulin. yG-Glob-ulin contributed 5 to 14%o, albumin 2 to 9%, andsiderophilin 0 to 2.5%o of the total protein. Inany one volunteer, the four measured componentsaccounted for between 31 and 73% of the totalprotein. The concentrations of -yG-globulin andsiderophilin observed in these specimens suggestthat, in relation to concentrations of albumin pres-ent, -yG-globulin and siderophilin accumulate se-lectively in nasal secretions.

Acknowledgments

The assistance of Dr. David W. Alling in the statisticalevaluation of the data and the competent technical assist-ance of Mr. Clarence Szwed are gratefully acknowledged.

References

1. Amoss, H. L., and E. Taylor. Neutralization of thevirus of poliomyelitis by nasal washings. J. exp.Med. 1917, 25, 507.

2. Howitt, B. F. Relationship between nasal and hu-moral antipoliomyelitic substances. J. infect. Dis.1937, 60, 113.

3. Bell, E. J. The relationship between the antipolio-myelitic properties of human nasopharyngeal se-cretions and blood serums. Amer. J. Hyg. 1948,47, 351.

4. Cate, T. R., R. D. Rossen, R. G. Douglas, Jr., W. T.Butler, and R. B. Couch. The role of nasal secre-tion and serum antibody in the rhinovirus commoncold. Amer. J. Epidemiol. In press.

5. Remington, J. S., K. L. Vosti, A. Lietze, and A. L.Zimmerman. Serum proteins and antibody activityin human nasal secretions. J. clin. Invest. 1964,43, 1613.

6. Rossen, R. D., W. T. Butler, T. R. Cate, C. F. Szwed,and R. B. Couch. Protein composition of nasal se-cretion during respiratory virus infection. Proc.Soc. exp. Biol. (N. Y.) 1965, 119, 1169.

7. Banks, P. A., J. A. Kasel, M. Huber, R. H. Alford,and V. Knight. Persistence of neutralizing anti-body in adult volunteers immunized with adenovirussoluble antigen. Proc. Soc. exp. Biol. (N. Y.)1966, 121, 240.

8. Kabat, E. A., and M. M. Mayer. Experimental Im-munochemistry, 2nd ed. Springfield, Ill., CharlesC Thomas, 1961, p. 559.

9. Feinberg, J. G. Identification, discrimination andquantification in Ouchterlony gel plates. Int. Arch.Allergy 1957, 11, 129.

10. Tomasi, T. B., Jr., E. M. Tan, A. Solomon, and R. A.Prendergast. Characteristics of an immune sys-tem common to certain external secretions. J. exp.Med. 1965, 121, 101.

775

ROSSEN, SCHADE, BUTLER, ANDKASEL

11. Rossen, R. D., W. T. Butler, W. E. Vannier, R. G.Douglas, Jr., R. B. Couch, and A. G. Steinberg.The sedimentation and antigenic properties ofproteins in nasal and other external secretions.Submitted for publication.

12. Sober, H. A., F. J. Gutter, M. M. Wyckoff, and E. A.Peterson. Chromatography of proteins. II. Frac-tionation of serum protein on anion-exchange cel-lulose. J. Amer. chem. Soc. 1956, 78, 756.

13. Pennell, R. B. Fractionation and isolation of puri-fied components by precipitation methods in ThePlasma Proteins, F. W. Putnam, Ed. New York,Academic Press, 1960, p. 35.

14. Woodworth, R. C., and A. L. Schade. Immunologi-cal precipitin titrations based on radioactive tag-ging of the iron naturally chelated by the proteinssiderophilin and conalbumin. Biochim. biophys.Acta (Amst.) 1961, 47, 403.

15. Dixon, W. J., and F. J. Massey, Jr. Introduction toStatistical Analysis, 2nd ed. New York, McGraw-Hill, 1957.

16. Siegel, S. Nonparametric Statistics for the Be-havioral Sciences. New York, McGraw-Hill, 1956.

17. Fahey, J. L., and M. E. Lawrence. Wuantitatve uc-termination of 6.6 s'-y-globulins, 02A-globulins andy1-macroglobulins in human serum. J. Immunol.1963, 91, 597.

18. Heide, K., and H. Haupt. Darstellung noch nichttherapeutisch angewandter Plasmaproteine. Beh-ringwerk-Mitteilungen 1964, 43, 161.

19. Vaerman, J. P., H. H. Fudenberg, L. B. Johnson,and W. J. Mandy. Size heterogeneity of normaland pathological y1A globulins. Fed. Proc. 1964,23, 558.

20. McFarlin, D. E., W. Strober, R. D. Wochner, andT. A. Waldmann. Immunoglobulin A productionin ataxia telangiectasia. Science 1965, 150, 1175.

21. Rutstein, D. D., E. F. Ingenito, and W. E. Reynolds.The determination of albumin in human bloodplasma and serum. A method based on the inter-action of albumin with an anionic dye-2- (4'-hy-droxybenzeneazo) benzoic acid. J. clin. Invest.1954, 33, 211.

22. Schade, A. L. The microbiological activity of sider-ophilin in Protides of the Biological Fluids, Pro-ceedings of the 8th Colloquium, Bruges, 1960, H.Peeters, Ed. Amsterdam, Elsevier, 1961, p. 261.

776