Stimulation with Carbachol Alters Endomembrane Distribution and Plasma Membrane Expression of Intracellular Proteins in Lacrimal Acinar Cells

Exp. Eye Res. (1999) 69, 651–661Article No. exer.1999.0742, available online at http :}}www.idealibrary.com on

Stimulation with Carbachol Alters Endomembrane Distribution

and Plasma Membrane Expression of Intracellular Proteins in

Lacrimal Acinar Cells

TAO YANGa, HONGTAO ZENGb, JIAN ZHANGc, CURTIS T. OKAMOTOb,

DWIGHT W. WARRENc, RICHARD L. WOODc, MICHAEL BACHMANNe

AUSTIN K. MIRCHEFFd

a Departments of Physiology & Biophysics, b Pharmaceutical Sciences and Schools of Medicine, and

Pharmacy, c Cell & Neurobiology, d Ophthalmology, University of Southern California, Los Angeles, CA,

U.S.A., e Institut fuX r Physiologische Chemie, Fachbereich Medzin Johannes Gutenberg-UniversitaX t Mainz,

Mainz, Germany

(Received Columbia 30 November 1998 and accepted in revised form 21 July 1999)

The events that lead to Sjo$ gren’s autoimmune processes in the lacrimal gland remain poorly understood.The acinar cell’s responses to acute cholinergic stimulation include release of secretory products acrossthe apical plasma membrane (apm) and a number of processes related to traffic between endomembranecompartments and the basal–lateral plasma membranes (blm), such as recruitment of Na,K-ATPase,accelerated recycling, and accelerated transcytosis of secretory IgA. We tested the hypothesis thatstimulation-induced acceleration of endomembrane traffic is accompanied by changes in compart-mentation and increased blm expression of proteins that are normally sequestered in endomembranecompartments. Isolated rabbit lacrimal gland acinar cells were cultured in serum-free media for 2 days.After harvesting, cells were incubated with or without 10 µ carbachol at 37°C for 20 min. Cells werelysed, and lysates were analysed by isopycnic centrifugation on sorbitol gradients. Galactosyltransferasecatalytic activity was determined biochemically. Different forms of cathepsin B were detected by Westernblotting. Carbachol stimulation decreased the contents of β-hexosaminidase, α-glucosidase, and proteinin secretory vesicles and increased them in specific compartments of the trans-Golgi network (ld-tgns).Stimulation also caused levels of galactosyltransferase, preprocathepsin B, and procathepsin B to increasetwo- to three-fold in the blm as well as increasing in the ld-tgns. Other changes caused by sustainedstimulation included: (a) increased levels of protein and procathepsin B in compartments of the lysosomalpathway; (b) changes in the distributions of Rab5 within the endomembrane system; (c) changes in thedistribution of Rab6 within the Golgi complex and tgn; (d) decreased expression of acid phosphatase andMHC class II molecules in the blm; and (e) decreased total content of Na,K-ATPase, which appeared tohave been selectively depleted from the tgn and blmre. We propose that the normal compartmentationof certain proteins may allow them to remain cryptic, such that they are not subject to central tolerance.Stimulation-induced increases in the levels expressed at the blm or secreted to the interstitium may,therefore, contribute to initiation of local autoimmune responses. # 1999 Academic Press

Key words : Sjo$ gren’s syndrome; autoimmune dacryoadenitis ; autoantigen expression; Golgi ;lysosomes; galactosyltransferase ; cathepsin B; membrane traffic.

1. Introduction

Autoimmune processes contribute to lacrimal gland

dysfunction and dry eye in Sjo$ gren’s syndrome and

possibly other conditions, such as diffuse infiltrative

lymphocytosis syndrome and cryptic autoimmune

dacryoadenitis (Mircheff et al., 1996). Autoimmunity

is considered to result from a failure of self-tolerance,

but neither the mechanisms that establish and main-

tain self-tolerance nor the reasons for their failure are

thoroughly understood. One attractive concept is that

of a cryptic self, comprised of autoantigens which are

not subject to tolerance because they are not presented

* Address correspondence to: Austin K. Mircheff, Department ofPhysiology & Biophysics, University of Southern California, Schoolof Medicine, 1333 San Pablo Street, MMR 626, Los Angeles,CA 90033, U.S.A.

to the immune system in sufficiently high concen-

trations (Gammon et al., 1991). Several theories have

been proposed to account for initiation of immune

responses to normally cryptic autoantigens. In mol-

ecular mimicry, lymphocytes are activated against

epitopes of viral and bacterial proteins, then react

against similar epitopes of self proteins (Oldstone,

1987; Walker and Jeffrey, 1986; Brusic et al., 1997).

In a different form of mimicry, occurring at the cellular

level, cells are aberrantly induced to express major

histocompatibility complex Class II molecules (MHC

II), then process and present tissue-specific auto-

antigens to reactive CD4 lymphocytes with mechan-

isms mimicking those of the professional antigen

presenting cells (Londei et al., 1984; Mircheff et al.,

1991). In a third mechanism, autoantigens which are

normally sequestered in endomembrane compart-

0014–4835}99}120651­11 $30.00}0 # 1999 Academic Press

652 T. YANG ET AL.

ments begin to be expressed in the plasma membranes,

where they may be recognized by B lymphocyte

antigen receptors, or released to the interstitium,

where they may be processed and presented by

professional antigen presenting cells (Lehmann et al.,

1992; Lanzavecchia, 1995). In putative examples of

this third mechanism, the ribonuclear proteins

Ro}SSA and La}SSB, frequent targets of autoimmune

responses in Sjo$ gren’s syndrome and systemic lupus

erythematosus (Alexander et al., 1982; Lee, 1988),

appear in the cytoplasm and plasma membranes of B

lymphocytes and epithelial cells that have been

infected with viruses or exposed to cytokines, and

in apoptotic bodies of irradiated keratinocytes

(Bachmann et al., 1989; Jones, 1992; Clark et al.,

1994).

The normal physiological function of lacrimal gland

acinar cells depends on a dynamic traffic between

endomembrane compartments and the surface mem-

branes. The apically-directed merocrine pathway for

synthesis, processing, storage, and release of secretory

proteins and recycling of secretory vesicle membranes

has been recognized for many years (e.g. Farquhar,

1981). Less widely appreciated traffic pathways are

directed to, and originate from, the basal–lateral

plasma membrane (blm). The transcytotic pathway

for dimeric IgA secretion involves endocytosis at the

blm and sequential transfers to the blm recycling

endosome (blmre), the Golgi complex (Gol) and one or

more specific domains of the trans-Golgi network (tgn)

(Zeng et al., 1998). Reserves of Na,K-ATPase pumps

are recruited to the basal–lateral membrane from the

tgn to maintain Na+ homeostasis following Na+}H+

antiporter activation during stimulated salt secretion

(Lambert et al., 1993b). Endocytosed fluid phase

markers reach the tgn, the Gol, and the lysosomal

pathway (Gierow et al., 1996). Preliminary studies in

our laboratories suggest that certain internalized

receptor–ligand complexes are targeted to the tgn,

while others are directed to the lysosomal pathway.

Vesicular traffic to the blm may mediate the surface

expression of La}SSB, once cytoplasmic expression of

this autoantigen has been induced. It could also

mediate the MHC II-restricted presentation of proteo-

lytically processed autoantigen peptides in cells which

have been induced to express MHC II (Yang et al.,

1999). Recent work suggests a possible additional way

traffic from endomembrane compartment to the

basal–lateral membrane might be involved in initiating

autoimmune responses. Like traffic in the apical

merocrine pathway, traffic related to the blm pathway

is influenced by secretomotor stimulation. The cholin-

ergic agonist, carbachol, stimulates release of secretory

component (Yoshino et al., 1995), and it accelerates

both endocytosis and exocytosis of surface (Lambert et

al., 1993a) and fluid phase markers (Gierow et al.,

1995). Carbachol also increases the flux of internalized

fluid phase markers into compartments where they

become sequestered (Gierow et al., 1995), and it

increases the appearance of β-hexosamindase in Golgi-

related compartments (Hamm-Alvarez et al., 1997).

In another glandular cell type, the acinar cell of the

exocrine pancreas, supramaximal stimulation per-

turbs segregation between the apical secretory path-

way and the lysosomal pathway, causing lysosomal

procathepsins to accumulate in the same compart-

ment as secretory zymogens (Saluja et al., 1987). It

has been postulated that the ensuing cathepsin-

mediated activation of secretory proteases causes

acinar cell autolysis and acute hemorrhagic pan-

creatitis. By analogy, it seems reasonable to hypothe-

size that sustained or supramaximal stimulation of

lacrimal gland acinar cells alters distributions of

intracellular proteins in endomembrane compart-

ments and increases levels at which they are expressed

in the basal–lateral membrane or secreted to the

interstitium.

In this study methods of analytical subcellular

fractionation were used to survey the effects of

stimulation with carbachol on the compartmentation

of selected intracellular proteins in reconstituted acini

from rabbit lacrimal glands. After an initial equi-

libration period the cells were exposed to carbachol at

a concentration of 10 µ for 20 min, then lysed and

analysed by differential and isopycnic centrifugation

on sorbitol density gradients. The 10 µ carbachol

concentration was selected because it has been found

to optimally accelerate recycling traffic to and from the

basal–lateral membrane while half-maximally stimu-

lating merocrine protein secretion across the apical

plasma membrane. The 20 min stimulation period

was selected because the rates of apical protein

secretion and of basal–lateral recycling traffic return

toward basal levels by this time, but the effects on

compartmentation of internalized fluid phase markers

remain evident (Gierow et al., 1995).

2. Materials and Methods

Materials

Female New Zealand white rabbits weighing 2 kg

were obtained from Irish Farms (Norco, CA, U.S.A.).

Anti-human cathepsin B polyclonal Ab, raised in

sheep, was from The Binding Site, Ltd. (Birmingham,

U.K.). Antibodies to Rab5 and Rab6, raised in rabbits

against synthetic peptides of human origin, were from

Santa Cruz Biotechnology Inc. (Santa Cruz, CA,

U.S.A.). Horseradish peroxidase Type II (HRP) and

Leupeptin were from Sigma Chemical Company (St.

Louis, MO, U.S.A.). "#&I-labeled protein A and "#&I-

labeled protein G were from ICN (Irvine, CA, U.S.A.).

Prestained SDS-PAGE standards, low range (19–

107 kD) and high range (47–205 kD) were purchased

from BioRad (Hercules, CA, U.S.A.). All other reagents

were purchased from standard suppliers as described

in previous reports (Gierow et al., 1996; Hamm-

Alvarez et al., 1997).

ALTERING THE LACRIMAL AUTOANTIGEN SPECTRUM 653

Isolation, Culture, and Incubation of Acinar Cells

Lacrimal gland acinar cells were isolated with the

same methods employed in previous studies (Gierow et

al., 1996; Hamm-Alvarez et al., 1997; Yang et al.,

1999). Cells were harvested after 40 hr incubation

and washed twice with culture medium. They were

resuspended in culture medium containing HRP at a

final concentration of 5 mg ml−" and leupeptin, a non-

cytotoxic thiol and serine protease inhibitor, at a final

concentration of 1 µg ml−". The cells were divided into

two 4 ml samples, each containing about 4¬10) cells,

and placed in 25 ml Erlenmeyer flasks. In some

experiments carbachol was added to one of the paired

flasks at a final concentration of 10 µ. In other

experiments both cell samples were incubated in the

absence of carbachol and later pooled for further

analysis. After 20 min at 37°C, the cell suspensions

were diluted by addition of 7 ml ice-cold Hanks’ buffer

and washed three times with 7 ml aliquots of the same

buffer. Washed cells were resuspended in 2 ml ice-cold

5% sorbitol cell lysis buffer containing protease

inhibitors, then lysed by 20 passages through a 20

gauge syringe needle followed by 30 passages through

a Balch cell press (H & Y Enterprises, Redwood City,

CA, U.S.A.).

Subcellular Fractionation

Subcellular fractionation procedures were those

described by Gierow et al. (1996) and Hamm-Alvarez

et al. (1997). The lysate was centrifuged for 10 min at

1000 g in a Sorvall RT 6000D centrifuge at 4°C. The

pellet was resuspended, homogenized and centrifuged

again. The combined supernatant fractions were

mixed with a pre-cooled solution of 87±4% sorbitol in

a 1±0:1±5 ratio, then loaded onto preformed sorbitol

density gradients. The density gradients were centri-

fuged at 100000 g for 5 hr at 4°C in a Beckman

SW28 rotor. The gradient contents were collected in

12 fractions, diluted with sorbitol cell lysis buffer, and

centrifuged at 250000 g for 90 min at 4°C. Resulting

pellets, as well as the pellets forming beneath the

density gradients, were resuspended in 1 ml aliquots of

the sorbitol cell lysis buffer and snap frozen in liquid

nitrogen prior to storage at ®80°C.

Analytical Methods

Protein and biochemical markers were determined

with methods that have been described in detail

elsewhere (Gierow et al., 1996; Hamm-Alvarez et al.,

1997; Yang et al., 1999). For immunoblotting, 20 µl

aliquots of each fraction were dissolved in denaturing

sample buffer by boiling for 5 min. SDS-PAGE was

carried out in the BioRad (Hercules, CA, U.S.A.)

Modular Mini-PROTEAN II apparatus. Proteins were

transferred to nitrocellulose membrane in Mini Trans-

Blot Cells (BioRad). Membranes were blocked, incu-

bated with primary antibodies, and labeled with ["#&I]-

protein A or ["#&I]-protein G. Washed membranes were

autoradiographed on Kodak XAR film at ®80°C.

Typical autoradiograms of immunoblots for markers

described in this study are presented elsewhere (Yang

et al., 1999). The films were scanned with a GS-670

Imaging Densitometer (BioRad) and analysed with

Molecular Analyst PC4 image analysis software.

As described in reports of previous fractionation

studies (Gierow et al., 1996; Hamm-Alvarez et al.,

1997; Yang et al., 1999), marker contents of

individual density gradient fractions were expressed as

percentages of the totals summed over the 13 fractions.

Summed activities were also normalized to the total

amount of membrane-associated protein summed over

the 13 fractions. Stimulation with carbachol did not

significantly change this parameter for any marker

except NaK-ATPase. Differences between marker con-

tents in corresponding fractions from resting and

carbachol-stimulated samples were evaluated with

Student’s t-test, with P!0±05 as the criterion for

significance.

3. Results

The low-speed supernatant fraction which was

analysed by isopycnic centrifugation contains micro-

somal populations derived from most of the cells

membrane-bounded organelles. Consequently, the

density gradient distributions of membrane markers

comprise low-resolution maps of intracellular bio-

F. 1. Summary of density gradient positions of principalmembrane compartments. Gol, Golgi complex; ld-tgns, low-density trans-Golgi network compartments ; hd-tgns, high-density trans-Golgi network compartments ; sv, secretoryvesicles ; preLys, pre-lysosomal (late endosomal) compart-ments ; Lys, lysosomes; blm, basal–lateral membranes;blmre, basal–lateral membrane recycling endosome.

654 T. YANG ET AL.

F. 2. Density gradient distributions of β-hexosaminidase, α-glucosidase, and protein in resting cells and cells that had beenstimulated with 10 µm carbachol for 20 min, and stimulation-induced changes. Values presented are mean percentages of totalrecovery activity³..(.) Values in parenthesis are numbers of preparations for which markers were analysed. Stimulationcaused all three markers to redistribute from the sv to ld-tgn. Total protein was additionally redistributed to the higher densitypreLys or the Lys.

F. 3. Density gradient distributions of galactosyltransferase. Details are as in Fig. 2. Stimulation increased galactosyl-transferase expression in the blm and domains of the ld-tgn.

chemical organization. Membranes from several endo-

membrane compartments often occupy overlapping

locations in the density gradient, but they can be

resolved from one another by partitioning in aqueous

Dextran-polyethyleneglycol phase systems (Mircheff,

1997). Maps of the two-dimensional, i.e. density vs

phase partitioning, distributions of the endomembrane

compartments that have been detected and charac-

terized in resting cells are presented elsewhere (Yang

et al., 1999). Fig. 1 summarizes the density gradient

positions of these compartments.

The density gradient distributions of β-hexosamini-

ALTERING THE LACRIMAL AUTOANTIGEN SPECTRUM 655

F. 4. Density gradient distributions of preprocathepsin B, procathepsin B, and cathepsin B. Details are as in Fig. 2.Stimulation increased expression of the prepro- and pro- forms in the blm and domains of the ld-tgn.

dase, α-glucosidase, and protein (Fig. 2) were similar

to each other in resting cells, and carbachol stimu-

lation caused roughly parallel changes in all three,

decreasing them significantly in fraction 11. The

recovery of β-hexosaminidase increased in fractions

2–4 and in fraction 6. The recoveries of α-glucosidase

and protein increased in fractions 2–6. These changes

indicate that stimulation caused membrane-associated

components of these markers to redistribute from the

secretory vesicles (sv) to low-density domains of the

trans-Golgi network, i.e. the ld-tgns. Additionally, the

recovery of protein increased significantly in fraction

P, indicating a redistribution to the higher density

preLys or to one or more of the Lys. In each case the

redistribution involved a small fraction of the total

recovered marker, but the increases represented

substantial changes relative to the amounts present in

the individual fractions in the resting state.

The density gradient distribution of galactosyl-

transferase (Fig. 3) was distinct from those of β-

hexosaminidase, α-glucosidase, and protein. However,

the general pattern of the stimulation-induced changes

was similar to those of the markers depicted in Fig. 2.

That is, galactosyltransferase recovery decreased sig-

nificantly in fraction 12 and increased in fractions 1,

2, 4 and 6. The change in fraction 1 represented a

three-fold increase in the amount expressed in the

blm. The compartments which lost galactosyltrans-

ferase could not be identified because of the degree of

variability between preparations, but the possible

candidates include the Gol, blmre, and high density

domains of the tgn (hd-tgns).

Fig. 4 summarizes the distributions of prepro-, pro-,

and mature forms of cathepsin B. The prepro- and

mature forms exhibited similar distributions in resting

cells. The pro form overlapped the prepro- and mature

forms, but was present at considerable excess in

fractions 3–5. Stimulation appeared to decrease

expression of all three forms in fractions 8–11,

although the decreases for the different forms were

statistically significant in different fractions. As the

cathepsin B isoforms are expressed in a number of

compartments that overlap between fractions 8 and

11, it is difficult to determine the precise compartments

from which the immature and mature form were

removed. Each form exhibited a distinct redistribution

656 T. YANG ET AL.

F. 5. Density gradient distributions of Rab6. Details are as in Fig. 2. Stimulation increased expression of Rab6 in the lg-tgn.

F. 6. Density gradient distributions of Rab5. Details are as in Fig. 2. Stimulation caused Rab5 to redistribute from the hd-tgn to the ld-tgn.

F. 7. Density gradient distributions of acid phosphatase and HRP, which was adsorbed to blm and internalized byendocytosis. Details are as in Fig. 2. Stimulation decreased acid phosphatase expression in the blm and increased HRP flux intothe ld-tgn.

ALTERING THE LACRIMAL AUTOANTIGEN SPECTRUM 657

F. 8. Density gradient distributions of Na,K-ATPase. Values presented are adjusted by the ratio of total recovered Na,K-ATPase activity to total recovered membrane protein. Other details are as in Fig. 2. Stimulation caused a 15³6% (P!0±03)decrease in the total activity, particularly affecting compartments of the ld-tgn and, presumably,higher density compartments of the blmre.

F. 9. Density gradient distributions of La}SSB and MHC II. Details are as in Fig. 2. Stimulation increased expression ofLa}SSB in the hd-tgn. It decreased expression of MHC II in the blm and compartments of the ld-tgn.

pattern. Expression of the prepro form and the pro

form increased in fraction 1 by factors of three and

two, respectively. The prepro form also increased in

fractions 3 and 4, while the pro form increased in

fraction 2 and in fraction P. Thus, stimulation

increased expression of prepro- and procathepsin B in

the blm and in separate ld-tgn microdomains. It

additionally caused procathepsin B to accumulate in

the higher density preLys or the Lys.

The distributions of Rab6 and Rab5 are summarized

in Figs 5 and 6. Recovery of Rab6 increased in

fractions 2–5 following stimulation, indicating that it

was redistributed within the tgn and Golgi complex,

from the Gol or hd-tgns to the ld-tgns. Recovery of

Rab5 decreased in fraction 12 and increased in

fractions 6 and 7, suggesting that it was redistributed

to ld-tgns or to lower density microdomains of the

blmre.

The distributions of acid phosphatase and HRP (Fig.

7) were similar to each other in resting cells. Acid

phosphatase decreased significantly in fraction 1,

indicating that it was removed from the blm. HRP

increased significantly in fractions 4 and 5, suggesting

relatively increased flux to domains of the ld-tgn.

Uniquely among the markers examined, the total

content of Na,K-ATPase (Fig. 8) decreased significantly

following stimulation, suggesting that stimulation

increased flux to Lys, where it was degraded. This

decrease was not distributed uniformly across all the

compartments. Rather, its impact was most pro-

658 T. YANG ET AL.

nounced in fractions 2 and 3 and in fractions 10 and

11, suggesting that blm expression was maintained at

the expense of pump units removed from domains of

the tgn and higher density domains of the blmre.

The distributions of La}SSB and MHC II are

summarized in Fig. 9. These markers overlapped in

fraction 1 and in fractions 7–P. Additionally, there

was an excess of MHC II over La}SSB in fractions 3–6.

Stimulation increased the content of La}SSB in

fraction 10, where it is primarily associated with the

hd-tgn. Similar to its effect on the distribution of acid

phosphatase, stimulation decreased the recovery of

MHC II in fraction 1 and in fraction 6, suggesting that

MHC II was redistributed away from the blm and

domains of the ld-tgns. However, the effects of

stimulation on MHC II recovery in the higher density

F. 10. Working hypothesis for organization of endomembrane traffic in the lacrimal gland acinar cell. White and grayarrows indicate postulated traffic pathways. Black arrows indicate likely sites of sorting processes that are altered during 20 minstimulation with carbachol. As discussed in the text, stimulation accelerates virtually every traffic step; it alters sorting in theld-tgn and blmre, and it accelerates traffic from preLys to Lys. Abbreviations are as in the legend to Fig. 1. Additionalcompartments which have not yet been resolved by analytical fractionation procedures are the apical membranes (apm), theapical endosome (ae), and the terminal transcytotic vesicles (ttv).

fractions varied between preparations, and it was not

possible to identify the destination of the MHC II

redistribution.

4. Discussion

Lacrimal gland acinar cells continually synthesize

secretory and lysosomal proteins, recycle blm constitu-

tents, and generate a transcytotic traffic of IgA

mediated by polymeric immunoglobulin receptors

(pIgR). Thus, many content and membrane proteins

will reside, at least transiently, in a multiplicity of

compartments of the biosynthetic–, apical secretory–,

lysosomal–, and basal–lateral membrane pathways.

The analytical approach we have used allows us to

ALTERING THE LACRIMAL AUTOANTIGEN SPECTRUM 659

infer the approximate positions isolated endomem-

brane compartments occupy in the sorbitol density

gradients. Despite the facts that specific proteins are

present in several different compartments, that the

compartments’ boundaries are often indistinct, and

that several different compartments overlap each other

in the density gradients, the data presented in Figs 2–9

indicate that it is feasible to detect changes in the

manner in which a number of proteins are distributed

among various compartments.

A further limitation of the analytical approach is

that the identities of the isolated compartments in

terms of defined structures of intact cells must be

regarded as working hypotheses, rather than definitive

conclusions. Our working hypotheses for the compart-

ments identities are summarized in the legend to Fig.

1, and our working hypothesis for organization of

traffic pathways between the compartments is depicted

in Fig. 10. The cellular model in Fig. 10 integrates data

from a wide variety of epithelial and non-epithelial

systems, as well as results of studies on membrane

traffic in lacrimal acinar cells (e.g. Gierow et al., 1996;

Hamm-Alvarez et al., 1997). The tgn plays a central

role, sorting proteins into transport vesicles that

mediate return to the Gol (van Weert et al., 1997) and

forward transit to various compartments (Keller and

Simons, 1997), including immature secretory vesicles

(isv) (Arvan and Castle, 1998), apical (apm) and blm

(Zegers and Hoekstra, 1998) and endosomes and pre-

lysosomes (preLys) (Hunziker and Geuze, 1996).

Sorting processes in the isv return lysosomal and other

proteins to the endomembrane system, although the

precise pathway is not yet clear (Klumperman et al.,

1998). Terminal transcytotic vesicles (ttv) are pre-

sumed to mediate the final step of pIgR transcytosis

(Zeng et al., 1998; Hansen et al., 1999). As many as

four pathways, each with its own populations of

transport vesicles, may mediate traffic between the

blmre and the blm, Gol (Llorente et al., 1998), tgn

(Nakajima and Pfeffer, 1997), and preLys. After

stimulated exocytosis, constituents of sv membranes

(svm) are retrieved and translocated, first, presumably,

to an apical endosome (ae) which communicates with

the Golgi (Farquhar, 1981) and with the Lys (Lerch et

al., 1995). In addition to accelerating exocytic fusion

of sv and recycling of svm constituents, secretomotor

stimulation accelerates both in-bound and out-bound

steps of the blm – blmre traffic, traffic to the ld-tgns,

and traffic to the Lys (Lambert et al., 1993a; Gierow et

al., 1995).

Yang et al. (1999) have recently summarized

observations supporting hypotheses for identities of

isolated compartments with structures depicted in Fig.

10. Identities of the compartments designated blm and

ld-tgn are most critical to the thesis that stimulation

with carbachol alters traffic of intracellular proteins in

ways that increase their expression at the blm or their

secretion to the interstitium. Both hypotheses are

based on the results of experiments in which HRP was

used as a marker for surface membranes and for

endocytosed fluid phase (Fig. 7 and Gierow et al.,

1996), and on the observed distribution (Fig. 5) of

Rab, which is known to be associated with the Golgi

complex and trans-Golgi network in a variety of cells

(Martinez et al., 1994).

Large components of the cell’s β-hexosaminidase

and α-glucosidase are contained in the cisternae of sv

and Lys (Hamm-Alvarez et al., 1997) and released to

the soluble phase upon cell lysis, when the intact

organelles are fragmented to form microsomal vesicles.

Thus, the high-speed supernatant fractions from

resting cells contain 75% of the α-glucosidase and

69% of the β-hexosaminidase. Significant fractions of

the smaller amounts that remain associated with the

membrane phase redistribute from svm to the ld-tgns

(Fig. 2) during 20 min stimulation. The path of this

redistribution is not clear. Secretory products re-

maining adsorbed to svm constituents may be returned

to the tgn after stimulated exocytosis. Additionally,

sorting in the tgn may be altered in ways that allow

increased amounts of newly synthesized and recycled

secretory proteins to reach the ld-tgns. The accumu-

lation of Rab6 in the ld-tgns (Fig. 5) accords with the

hypothesis that sorting in the tgn is altered during

20 min stimulation with carbachol.

As depicted in Fig. 10, galactosyltransferase and the

various forms of cathepsin B enter the tgn and the

blmre. In unstimulated cells, most of the galactosyl-

transferase and cathepsins entering the blmre are

retrieved to the Gol and other endomembrane compart-

ments, and only small amounts are allowed to move

forward to the blm. Cathepsin B, after some cycling

through the Gol, tgn, blmre, and preLys, is captured in

the Lys. Alterations of sorting within the tgn are likely

to account for the accumulation of galactosyltrans-

ferase (Fig. 3), preprocathepsin B, and procathepsin B

(Fig. 4) in the ld-tgn after 20 min.

The redistribution of Rab5 to the ld-tgns or to lower-

density domains of the blmre (Fig. 6) is indicative of

sorting changes within the endomembrane system.

Since Rab5 may be associated with secretory vesicles

(Wagner et al., 1994) and both the apical and the

basal-lateral endocytic pathways (Bucci et al., 1994),

the sites of Rab5 sorting cannot be discerned without

higher resolution fractionation data. The hypothesis

that sorting is altered in the blmre could plausibly

account for increased expression of galactosyltrans-

ferase, preprocathepsin B, and procathepsin B, and

for decreased expression of acid phosphatase and MHC

II, in the blm.

The increased amounts of procathepsin B and total

protein in the Lys after 20 min stimulation accord

with the observation that 10 µ carbachol increases

accumulation of Lucifer Yellow, a fluid phase marker,

in a sequestered compartment presumed to be the

lysosomes (Gierow et al., 1995). The net decrease in

the cells’ total content of Na,K-ATPase (Fig. 8) is also

consistent with increased traffic to the Lys.

660 T. YANG ET AL.

There is no need to postulate that cathepsins,

galactosyltransferase, α-glucosidase and β-hexosamin-

idase are specifically targeted in autoimmune dacryo-

adenitis. However, we propose that the changes in

their distributions occurring during 20 min stimu-

lation are emblematic of changes of protein compart-

mentation which might, potentially, challenge per-

ipheral tolerance. There are two- to three-fold in-

creases in the levels at which galactosyltransferase

and cathepsins are expressed in the blm, where

proteins can be detected by surface antigen receptors

of reactive B cells. Since the ld-tgns communicate with

the blm, the significant increases of α-glucosidase and

β-hexosaminidase in the ld-tgns could reflect increases

in the levels at which some soluble proteins are se-

creted to the interstitium, where they are subject to

internalization, processing, and presentation by pro-

fessional antigen presenting cells. In acinar cells which

express MHC II, changes of compartmentation would

change the spectrum of MHC II-associated peptides

presented at the blm and undermine peripheral

tolerance (Mircheff et al., 1994).

The thesis that stimulation-induced increases in blm

expression and secretion of normally cryptic auto-

antigens contribute to the initiation of autoimmune

dacryoadenitis may help explain why Sjo$ gren’s syn-

drome occurs most frequently in peri- and post-

menopausal women. The lacrimal glands atrophy

significantly in women between the ages of 20 and 70

years (Ueno et al., 1996), presumably due to loss of

androgenic support (Mircheff et al., 1996; Azzarolo et

al., 1997). The lacrimal gland is the effector of a

physiological system that keeps the ocular surface

well-lubricated and free of irritants. When lacrimal

gland secretory capacity decreases, this servomechan-

ism would be predicted to increase or prolong

secretomotor output to surviving tissue in response to

ocular surface irritation.

Lacrimal gland acinar cells share fundamental

mechanisms with acinar cells of the salivary glands.

Thyroid epithelial cells carry out regulated fluxes

along both the apical pathway, i.e. toward the

follicular lumen, and the basal–lateral pathway, i.e.

toward the interstitium. Pancreatic β-cells are not

epithelial, but they are engaged in the regulated

secretion of insulin. These tissues are all targets of

autoimmune responses, and it seems possible that

activation-induced alterations of basic cellular mecha-

nisms for sorting proteins into endomembrane traffic

pathways might influence the etiopathogenesis of

other autoimmune diseases in addition to Sjo$ gren’s

syndrome, including thyroiditis and Type I diabetes

mellitus.

Acknowledgements

The authors thank Hannah Freed, John M. Norian,Barbara W. Platler, Wei Wang, and Ramona Yasharel fortheir help completing this work. They thank Drs Sarah F.Hamm-Alvarez, Harvey R. Kaslow, Joel E. Schechter, and

Hermann von Grafenstein for advice and helpful discussions.Supported by NIH grants EY 05801 (AKM), EY 09405(DWW), and EY 10550 (RLW), Digestive Diseases CoreCenter Grant DK 48522, and a grant from the University ofSouthern California Zumberge Faculty Research and Inno-vation Fund (CTO).

References

Alexander, E. L., Hirsch, T. J., Arnett, F. C., Provost, T. T.and Stevens, M. B. (1982). Ro (SSA) and La (SSB)antibodies in the clinical spectrum of Sjo$ gren’s syn-drome. J. Rheumatol. 9, 239–46.

Arvan, P. and Castle, D. (1998). Sorting and storage duringsecretory granule biogenesis : looking backward andlooking forward. Biochem. J. 332, 593–610.

Azzarolo, A. M., Mircheff, A. K., Kaswan, R. L., Stancyzk,F. Z., Gentschein, E., Becker, L., Nassir, B. and Warren,D. W. (1997). Androgen support of lacrimal gland func-tion. Endocrine 6, 39–45.

Bachmann, M., Falke, D., Schroder, H. C. and Muller, W. E.(1989). Intracellular distribution of the La antigen inCV-1 cells after herpes simplex virus type 1 infectioncompared with the localization of U small nuclearribonucleoprotein particles. J. Gen. Virol. 70, 881–91.

Brusic, V., Rudy, G. and Harrison, L. C. (1997). Molecularmimicry – from hypothesis towards evidence. Immunol.Today 18, 95–6.

Bucci, C., Wandinger-Ness, A., Lutcke, A., Chiariello, M.,Bruni, C. B. and Zerial, M. (1994). Rab5a is a commoncomponent of the apical and basolateral endocyticmachinery in polarized epithelial cells. Proc. Natl. Acad.Sci. U.S.A. 91, 5061–5.

Clark, D. A., Lamey, P. J., Jarrett, R. F. and Onions, D. E.(1994). A model to study viral and cytokines in-volvement in Sjo$ gren’s syndrome. Autoimmunity 18,7–14.

Farquhar, M. G. (1981). Membrane recycling in secretorycells : implications for traffic of products and specializedmembrane within the Golgi complex. Methods Cell Biol.23, 399–427.

Gammon, G., Sercarz, E. E. and Benichou, G. (1991). Thedominant self and the cryptic self : shaping the auto-reactive T-cell repertoire. Immunol. Today 12, 193–5.

Gierow, J. P., Lambert, R. W. and Mircheff, A. K. (1995).Fluid phase endocytosis by isolated rabbit lacrimalgland acinar cells. Exp. Eye Res. 60, 511–25.

Gierow, J. P., Yang, T., Bekmezian, A., Liu, N., Norian, J.,Kim, S. A., Rafisolyman, S., Zeng, H., Okomoto, C. T.,Wood, R. L. and Mircheff, A. K. (1996). Na-K-ATPase inlacrimal gland acinar cell endosomal system: correctinga case of mistaken identity. Am. J. Physiol. 271,C1685–98.

Hamm-Alvarez, S. F., da Costa, S., Yang, T., Wei, X., Gierow,J. P. and Mircheff, A. K. (1997). Cholinergic stimulationof lacrimal acinar cells promotes redistribution ofmembrane-associated kinesin and the secretory protein,beta-hexosaminidase, and increases kinesin motoractivity. Exp. Eye Res. 64, 141–56.

Hansen, G. H., Niels-Christiansen, L. L., Immerdal, L.,Hunziker, W., Kenny, A. J. and Danielson, E. M. (1999).Transcytosis of immunoglobulin A in the mouseenterocyte occurs through glycolipid raft- and rab17-containing compartments. Gastroenterol. 116, 610–22.

Hunziker, W. and Geuze, H. J. (1996). Intracellular traf-ficking of lysosomal membrane proteins. Bioessays 18,379–89.

Jones, S. K. (1992). Ultraviolet radiation (UVR) induces cell-surface Ro}SSA antigen expression by human keratino-

ALTERING THE LACRIMAL AUTOANTIGEN SPECTRUM 661

cytes in vitro : a possible mechanism for the UVRinduction of cutaneous lupus lesions. Br. J. Dermatol.126, 546–53.

Keller, P. and Simons, K. (1997). Post-Golgi biosynthetictrafficking. J. Cell Sci. 110, 3001–9.

Klumperman, J., Kuliawat, R., Griffith, J. M., Geuze, H. J. andArvan, P. (1998). Mannose-6-phosphate receptors aresorted from immature secretory vesicles via adaptorprotein AP-1, clathrin, and syntaxin-6 positive vesicles.J. Cell Biol. 141, 359–71.

Lambert, R. W., Maves, C. A., Gierow, J. P., Wood, R. L. andMircheff, A. K. (1993a). Plasma membrane internaliz-ation and recycling in rabbit lacrimal acinar cells.Invest. Ophthalmol. Vis. Sci. 34, 305–16.

Lambert, R. W., Maves, C. A. and Mircheff, A. K. (1993b).Carbachol-induced increase of Na+}H+ antiport andrecruitment of Na+}K+-ATPase in rabbit lacrimal acini.Curr. Eye Res. 12, 539–51.

Lanzavecchia, A. (1995). How can cryptic epitopes triggerautoimmunity? J. Exp. Med. 181, 1945–8.

Lee, L. A. (1988). Anti-Ro (SSA) and anti-La (SSB) antibodiesin lupus erythematosus and in Sjo$ gren’s syndrome.Arch Dermatol. 124, 61–2.

Lehmann, P. V., Forsthuber, T., Miller, A. and Sercarz, E. E.(1992). Spreading of T-cell antoimmunity to crypticdeterminants of an autoantigen. Nature 358, 155–7.

Lerch, M. M., Saluja, A. K., Ru$ nzi, M., Dawra, R. and Steer,M. L. (1995). Luminal endocytosis and intracellulartargeting by acinar cells during early biliary pancratitisin the opossum. J. Clin. Invest. 95, 2222–31.

Llorente, A., van Deurs, B., Sandvig, K. (1998). Transport ofapically but not basolaterally internalized ricin to theGolgi is stimulated by 8-Br-cAMP in MDCK cells. FEBSLett. 431, 200–204.

Londei, M., Lamb, J. R., Bottazzo, G. F. and Feldmann, M.(1984). Epithelial cells expressing aberrant MHC class IIdeterminants can present antigen to cloned human Tcells. Nature 312, 639–41.

Martinez, O., Schmidt, A., Salame! ro, J., Hoflack, B., Roa, M.and Goud, B. (1994). The small GTP-binding proteinrab6 functions in intra-Golgi transport. J. Cell Biol. 127,1575–88.

Mircheff, A. K., Gierow, J. P., Lee, L. M., Lambert, R. W.,Akashi, R. H. and Hofman, F. M. (1991). Class IIantigen expression by lacrimal epithelial cells. Anupdated working hypothesis for antigen presentationby epithelial cells. Invest Ophthalmol Vis. Sci. 32,2302–10.

Mircheff, A. K., Warren, D. W. and Wood, R. L. (1996).Hormonal support of lacrimal function, primary lacri-mal deficiency, autoimmunity, and peripheral tolerance

in the lacrimal gland. Ocular Immunology and Inflam-mation. 4, 145–72.

Mircheff, A. K., Wood, R. L. and Gierow, J. P. (1994). Trafficof major histocompatibility complex Class II moleculesin rabbit lacrimal gland acinar cells. Invest. Ophthalmol.Vis. Sci. 35, 3943–51.

Mircheff, A. K. (1997). Methods and experimental analysisof isolated epithelial cell membranes. In: Epithelialtransport : a guide to methods and experimentalanalysis. Wills, N. K., Reuss, L. and Lewis, S. A. (Eds)Chapman and Hall, London, pp. 190–211.

Nakajima, Y. and Pfeffer, S. (1997). Phosphatidylinositol 3-kinase is not required for recycling of mannose 6-phosphate receptors from late endosomes to the trans-Golgi network. Mol. Biol. Cell 8, 577–82.

Oldstone, M. B. (1987). Molecular mimicry and autoimmunedisease. Cell 50, 819–20.

Saluja, A., Hashimoto, S., Saluja, M., Powers, R. E.,Meldolesi, J. and Steer, M. (1987). Subcellular re-distribution of lysosomal enzymes during caerulein-induce pancreatitis. Am. J. Physiol. 253, G508–16.

Ueno, H., Ariji, E., Izumi, M., Uetani, M., Hayashi, K. andNakamura, T. (1996). MR imaging of the lacrimalgland. Age-related and gender-dependent changes insize and structure. Acta Radiologica 37, 714–9.

Van Weert, A. W., Geuze, H. J. and Stoorvogel, W. (1997).Heterogeneous behavior of cells with respect to in-duction of retrograde transport from the trans-Golginetwork to the Golgi upon inhibition of the vacuolarproton pump. Eur. J. Cell Biol. 74, 417–23.

Wagner, A. C. C., Strowski, M. Z. and Williams, J. A. (1994).Identification of Rab5 but not Rab3A in rat pancreaticzymogen granule membranes. Biochem. Biophys. Res.Commun. 200, 542–8.

Walker, E. J. and Jeffrey, P. D. (1986). Polymyositis andmolecular mimicry, a mechanism of autoimmunity.Lancet 8507, 605–607.

Yoshino, K., Tseng, S. C. and Pflugfelder, S. C. (1995).Substrate modulation of morphology. Growth, and tearprotein production by cultured human lacrimal epi-thelial cells. Exp. Cell Res. 220, 138–51.

Yang, T., Zeng, H., Zhang, J., Okamoto, C. T., Warren, D. W.,Wood, R. L., Bachmann, M. and Mircheff, A. K. (1999).Major histocompatibility complex Class II molecules,cathepsins, and La}SSB proteins in lacrimal acinar cellendomembranes. Am. J. Physiol., in press.

Zegers, M. M. P. and Hoekstra, D. (1998). Mechanisms andfunctional features of polarized membrane traffic inepithelial and hepatic cells. Biochem. J. 336, 257–69.

Zeng, H., Qian, L., Yang, T., Mircheff, A. K. and Okamoto,C. T. (1998). The transcytotic pathway of dIgA in lac-rimal acinar cells. Invest. Ophthalmol. Vis. Sci. 39, S340.