Macrophage activation for tumor cell killing

MISSING IMAGE

Material Information

Title:
Macrophage activation for tumor cell killing analysis by two-dimensional gel electrophoresis
Physical Description:
vii, 138 leaves : ill. ; 29 cm.
Language:
English
Creator:
MacKay, Robert J., 1953-
Publication Date:

Subjects

Subjects / Keywords:
Macrophage Activation   ( mesh )
Cytotoxins -- therapeutic use   ( mesh )
Electrophoresis -- methods   ( mesh )
Pharmacy thesis Ph.D   ( mesh )
Dissertations, Academic -- Pharmacy -- UF   ( mesh )
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph.D.)--University of Florida, 1987.
Bibliography:
Bibliography: leaves 121-137.
Statement of Responsibility:
by Robert J. MacKay.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000901806
oclc - 17643470
notis - AEL0648
System ID:
AA00009095:00001


This item is only available as the following downloads:


Full Text














MACROPHAGE ACTIVATION FOR
TUMOR CELL KILLING: ANALYSIS BY
TWO-DIMENSIONAL GEL ELECTROPHORESIS





BY

ROBERT J. MACKAY


A DISSERTATION PRESENTED TO THE GRADUATE
SCHOOL OF THE UNIVERSITY OF FLORIDA IN
PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR
THE DEGREE OF DOCTOR OF PHILOSOPHY



UNIVERSITY OF FLORIDA


1987









ACKNOWLEDGEMENTS


I offer my most sincere appreciation to Dr. Steve

Russell for all his efforts on my behalf. I wish to thank

him for offering me a position in his laboratory, and for

all his encouragement and generosity. I have gained much

from him as both friend and mentor.

I also want to thank my committee members, Drs.

Feldherr, Normann, Johnson and Esser for their help and

guidance.

I have enjoyed working with all the members of Dr.

Russell's laboratory and especially wish to thank Sergio

and Kim for their expert technical assistance, Mark, Mike,

and Carol for being fellow sufferers and friends, and Drs.

Judy Pace and Mitali Basu for their helpful advice and

discussion. Finally, and most importantly, thanks to my

wife, Sally, for putting up with me and supporting me, for

without her this work would not have been completed.
















TABLE OF CONTENTS


ACKNOWLEDGEMENTS ............................... .. ii

ABSTRACT........................ .................... v

CHAPTERS

I INTRODUCTION

Historical Perspectives and Introduction
to Macrophage Activation............... 1
Requirements for In Vitro-Activation of
Macrophages for Tumor Cell Killing..... 4
Stepwise Activation of Macrophages for
Tumor Cell Killing..................... 6
Molecular Basis for Macrophage Activation
for Tumor Cell Killing................. 9
Molecular Markers for Stages of Macrophage
Activation............................. 19
Summary................................... 26

II PROTEIN CHANGES ASSOCIATED WITH STAGES
OF ACTIVATION OF BONE MARROW CULTURE-
DERIVED MACROPHAGES

Introduction.............................. 28
Materials and Methods ..................... 29
Results.... ............................. 36
Discussion................................ 52

III PROTEIN PHENOTYPES OF MOUSE MACROPHAGES
ACTIVATED IN VIVO FOR TUMOR CELL
KILLING

Introduction.............................. 57
Materials and Methods..................... 58
Results ................................... 62
Discussion................................ 75


iii










Page
IV MACROPHAGE ACTIVATION-ASSOCIATED PROTEINS:
RELATION OF PROTEIN SYNTHESIS TO PRIMING;
EXPRESSION OF PROTEINS 47B AND 71/73 IN
MACROPHAGES EXPOSED TO VARIOUS PRIMING
AND TRIGGERING AGENTS

Introduction..... ........... 82
Materials and Methods..................... 84
Results................................... 89
Discussion................................ 107

V SUMMARY OF RESULTS, CONCLUSIONS AND FUTURE
DIRECTIONS

Summary of Results........................ 113
General Conclusions and Future
Directions ............................. 117

REFERENCES......................................... 121

BIOGRAPHICAL SKETCH................................ 138















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment
of the Requirements for the Degree of
Doctor of Philosophy


MACROPHAGE ACTIVATION FOR
TUMOR CELL KILLING: ANALYSIS BY
TWO-DIMENSIONAL GEL ELECTROPHORESIS

By

Robert J. MacKay

May, 1987

Chairman: Stephen W. Russell
Cochairman: Alfred F. Esser
Major Department: Pathology

Two-dimensional gel electrophoresis was used to

examine the cellular protein changes associated with

macrophage activation for tumor cell killing. In

unstimulated bone marrow culture-derived (BMCD)

macrophages, approximately 2,000 proteins were resolved in

each gel. Forty proteins were induced in macrophages

activated by gamma interferon (IFN), a priming agent, plus

lipopolysaccharide (LPS), a triggering agent. Of these,

26 were induced by gamma IFN and 35 by LPS; 25 of the

LPS-induced changes were actually due to the autocrine

actions of secreted alpha/beta IFN. Two of the

activation-induced proteins, designated p47b and p71/73,









could be used to construct phenotypes for the stages of

macrophage activation: unstimulated macrophages expressed

neither protein, primed cells expressed p47b, and

activated, tumoricidal macrophages expressed both p47b and

p71/73. Peritoneal macrophages primed or activated

endogenously expressed the same protein phenotypes as

their BMCD macrophage counterparts. Furthermore, among

activated macrophages from either in vitro or in vivo

sources, there was a very close quantitative relation

between the expression of these proteins and cytolytic

activity. Beta IFN and several alternative triggering

agents also induced in BMCD macrophages the expression of

p47b and p71/73, respectively. Macrophages activated by

high concentrations of LPS alone expressed both proteins.

Resident peritoneal macrophages, like triggering

agent-treated BMCD macrophages, were p71/73 positive,

whereas inflammatory peritoneal macrophages expressed

neither p47b nor p71/73. Finally, cycloheximide-induced

inhibition of protein synthesis during exposure of

macrophages to IFN did not prevent the expression of p47b;

rather, its synthesis was deferred until protein synthesis

was resumed following the removal of cycloheximide. This

observation may explain reports, confirmed here, that

inhibition of protein synthesis had no effect on priming

of macrophages for tumor cell killing.









Taken collectively, these results indicate that the

presence or absence of proteins 47b and 71/73 can be used

to assess the activation status of macrophages, regardless

of the macrophage source or the activating stimulus used.

By using antibodies against these marker proteins, it

should be possible, for the first time, to examine

macrophage activation at the single cell level, either in

culture or in vivo in tumors or other lesions.


vii















CHAPTER I

INTRODUCTION
Historical Perspectives and Introduction to
Macrophage Activation
During the more than 100 years which have passed

since Elie Metchnikoff first described the activities of

phagocytic cells at an inflammatory site (reviewed in 1),

mononuclear phagocytes have been ascribed pivotal roles in

many aspects of homeostasis and defense against disease

(reviewed in 2). Macrophages, the most differentiated

cells in the mononuclear phagocytic series, are resident

normally in most body tissues and, in response to

inflammation, a large reserve of young, responsive

mononuclear phagocytes can be mobilized rapidly from the

bone marrow via the blood. Once exposed to the

inflammatory milieu, macrophages may acquire enhanced

competence to mediate various functions. The term

"activated" was introduced in the early 1960's by

Mackaness (reviewed in 3) to describe macrophages that had

been stimulated to a state of heightened physiological

activity characterized by changes in morphology and

function ("functional hypertrophy"). These changes

included cellular enlargement, increased pinocytic and









phagocytic activity, elevated content of hydrolytic

enzymes, organellar changes, increased spreading on glass,

and enhanced microbicidal ability. On the basis of his

classic studies of macrophages harvested from mice

chronically infected with intracellular pathogens,

Mackaness refined this definition and described the

general features of macrophage activation as they are

still understood today. He showed that activation during

chronic infection was immunologically controlled with an

induction phase that was antigen-specific and depended

upon the presence of lymphocytes which, Mackaness

speculated, acted on macrophages via a secreted soluble

mediator (4). Once activated, however, macrophages were

able to act nonspecifically. For example, macrophages

harvested from mice undergoing an immune response to

Listeria monocytogenes showed enhanced killing capacity

not only against the inducing bacteria, but also against

unrelated species such as Salmonella typhimurium (5).

Activation of macrophages, as measured by enhanced

bactericidal activity, was shown by Mackaness and his

colleagues to be evanescent and reversible, thereby

distinguishing the phenomenon from processes of

differentiation (6).

Activated macrophages were not only associated with

defense against infectious diseases; macrophages with

enhanced microbicidal activity were also found at sites of









noninfectious immunologic reactions such as allograft

rejection or graft-versus-host reaction. In addition,

macrophages activated during chronic infections were shown

to possess novel functional capacities apparently

unrelated to the enhanced killing of microbial pathogens.

Of particular relevance to this review were the findings

of Hibbs et al. (7) who showed, in 1972, using peritoneal

cells from mice chronically infected with Toxoplasma

gondii, that activated macrophages could efficiently lyse

tumor cells in vitro. Subsequent investigations showed

that infection with any of a variety of phylogenetically

unrelated organisms including Besnoitia jellisoni (8),

Listeria monocytogenes, or bacillus Calmette Guerin (9);

or vaccination with either Freund's complete adjuvant (9)

or killed C. parvum (10), induced in mice a population of

activated peritoneal macrophages that were tumoricidal

when tested in vitro. The cytolytic effect of activated

macrophages was shown to be selective for neoplastic cell

lines (7). The viability and growth of nonneoplastic

cells was not affected, even when these normal cells were

mixed with cultures of susceptible tumor cells (11, 12).

Elegant cinemicrographic studies by Meltzer et al. (12)

revealed that the presence of tumor cells greatly

stimulated the translational movement of activated

macrophages in culture. Intimate contact between

macrophages and tumor cells followed this rapid movement








and always preceded target cell destruction. Early

studies by Hibbs and his colleagues (7, 9) showed that

macrophage-mediated killing of tumor cells was a

nonphagocytic process requiring neither antibody nor

complement components. More recent investigations of the

killing mechanisms of activated macrophages will be

discussed in detail later in this review.



Requirements for In Vitro-Activation of
Macrophages for Tumor Cell Killing

Several early studies indicated that activation of

macrophages in vivo required the interaction of

specifically sensitized T lymphocytes with appropriate

antigen (4, 13). The link between specific cell-mediated

immunity and macrophage activation was further established

when Piessens et al. (14) and others (15, 16) showed that

the supernates from cultures of antigen- or mitogen-

simulated T lymphocytes could induce macrophages to become

cytolytic in vitro. The relevant lymphokine, termed

macrophage activating factor (MAF) by Piessens or

macrophage cytotoxicity factor (MCF) by Lohman-Matthes

(16), was generally associated with protein of approximate

Mr 50 kDa when lymphocyte supernates were fractionated by

gel filtration chromatography (17, 18). More recently,

biochemical and immunologic techniques in many

laboratories have strongly indicated that this activity is









mediated by gamma interferon (IFN) (19, 20). In

confirmation of these observations, purified gamma IFN,

produced by recombinant DNA technology, was shown to be a

potent MAF (21, 22, 23). The molecular identities of

other non-interferon MAF, including molecules of Mr 23 kDa

(24) and 30 kDa (25), remain to be established, although

the possibility that these are fragments of gamma IFN has

not been definitively excluded.

At the same time as the potent activating effect of

lymphokine was being revealed, several substances with no

apparent relevance to cell-mediated immunity also were

shown to be able to activate macrophages for tumor cell

killing. Among these were a group of anionic polymers

including double-stranded RNA (26), pyran copolymer, and

dextran sulfate (27), and a number of microbes or

microbical products or components such as endotoxin (26),

Brucella abortus ether extract (28), and muramyl dipeptide

(29). Schultz and his colleagues recognized that these

apparently unrelated substances could each induce the

secretion of alpha/beta IFN from a variety of cell types,

including macrophages. After showing that partially

purified alpha/beta IFN from L929 cells, like lymphokine,

could activate macrophages for tumor cell killing (30),

Schultz used a specific antiserum against alpha/beta IFN

to demonstrate that activation by polyanions and microbial

products was actually mediated by the autocrine actions of









alpha/beta IFN secreted by macrophages (31). These

observations have since been confirmed by others (32, 33).

Alpha/beta IFN secreted by macrophages may also mediate

the reported activating effects of colony stimulating

factor (CSF) (34). In support of this possibility, Moore

et al. (35) have clearly demonstrated that macrophages

exposed to purified CSF are stimulated to secrete IFN (and

several other cytokines, including plasminogen activator).

The issue of whether CSF activates macrophages directly,

which would be a novel finding, or via externalized IFN,

as was the case for endotoxin, should be resolved in the

near future by the use of specific anti-IFN antibodies.



Stepwise Activation of Macrophages
for Tumor Cell Killing

A major conceptual advance in the study of macrophage

activation was the discovery, in 1977, by Russell et al.

(36) and Hibbs et al. (37), working independently, that

there existed a stage of activation, exemplified by

macrophages isolated either from progressing Moloney

sarcomas (Russell), or from the peritoneal cavities of

BCG-infected mice (Hibbs), that was constitutively

noncytolytic but could be stimulated ("triggered" in the

terminology of Hibbs) to become tumoricidal by sub-ng/ml

concentrations of LPS; at least 100-fold higher

concentrations of LPS were required to activate









unstimulated macrophages. Thus, a new stage of activation

was placed between the unstimulated (i.e., resident or

bone marrow culture-derived macrophages) and fully

activated, cytolytic stages; these macrophages were termed

"activated noncytolytic" by Hibbs and "primed" by Russell.

Weinberg et al. (38) further clarified this concept of

stepwise activation by showing that LPS, as native

endotoxin, occurred ubiquitously in the environment and

contaminated most laboratory reagents, media, and

glassware. When steps were taken to rigorously exclude

this environmental endotoxin, lymphokine generally was

shown to prime, but not fully activate, macrophages for

tumor cell killing (39), providing an in vitro counterpart

to the tumor-derived population recognized by Russell et

al. and Hibbs et al. Like the latter macrophages,

lymphokine-primed cells were rendered fully cytolytic by

sub-ng/ml amounts of LPS or lipid A. Subsequent work,

using purified reagents, has demonstrated conclusively

that gamma IFN is the most important lymphokine that

primes (40). Alpha and/or beta IFN (type I IFN) also

could prime macrophages for tumor cell killing but, when

compared on the basis of antiviral activity, they were

500- to 1000-fold less efficient as priming agents than

gamma IFN was (40). In addition to LPS (or endotoxin), a

wide variety of microbial and nonmicrobial substances have

been reported to function effectively as triggering agents









in vitro (31, 32, 41, 42, 43). For certain highly

responsive populations of macrophages, concentrated

lymphokine preparations (44) have been shown to supply a

triggering signal. For macrophages that are fully

activated in vivo by processes unrelated to microbial

infection, e.g., in regressing Moloney sarcomas (45), the

source of endogenous triggering signals is presently

unknown.

According to Hibbs et al. (37), the unstimulated,

primed, and activated stages are not discrete entities but

are segments of a continuum that constitute the activation

process. Between the unstimulated (resident) and primed

stages of activation, Hibbs, and later Ruco and Meltzer

(46, 47) have inserted on additional stage or segment.

Macrophages stimulated to this stage, typically by the

intraperitoneal injection of proteose peptone or some

other sterile irritant, have been termed "responsive"

because they are more sensitive than resident macrophages

to the activating effects of lymphokine plus LPS; however,

unlike primed cells they are not rendered tumoricidal by

ng/ml amounts of LPS. A postcytolytic stage of activation

has also been identified. Although there has been general

agreement with Mackaness that full activation is a

short-lived phenomenon, reports of the functional

properties of postcytolytic macrophages have differed

markedly. For example, cytolytic macrophages isolated









from regressing Moloney sarcomas rapidly lost the ability

to kill when cultured in vitro, but retained (for at least

96 hr) sensitivity to trace amounts of LPS, i.e., they

were functionally primed (45). In contrast, resident

macrophages that were exposed in vitro to lymphokine plus

LPS lost cytolytic activity during 20 hr in culture and

could not be restimulated (48). These studies appear to

illustrate two fundamentally different processes:

modulation of function and differentiation/maturation,

respectively. It remains to be seen which postcytolytic

stage will prove to be more relevant to the process of

activation as it occurs naturally (i.e., in vivo).



Molecular Basis for Macrophage Activation
and Tumor Cell Killing

Receptors for Activating Agents

Recent studies have provided evidence for a surface

receptor for gamma IFN on mouse macrophages (49, 50).

Although a formal link between receptor occupancy and

elicited function has yet to be established, correlative

evidence indicates that gamma IFN must bind to specific,

saturable membrane structures on macrophages in order to

induce either priming for tumor cell killing (49) or cell

surface expression of Ia (50). Celada et al. (49)

reported that gamma IFN bound to a single class of sites

on macrophages with a dissociation constant (Kd) of 1 x









10-8M; an affinity significantly lower than the Kd of

1x10-10 to 1x109M reported for various nonmacrophage cell

lines (51, 52, 53). Aiyer et al. (50) recently provided

data which suggest that a separate class of high affinity

binding sites exists on macrophages, and argued that the

density (500 sites/cell) and affinity characteristics (Kd

= 9.1 x 10- 11M) of this putative receptor are consistent

with the very low concentrations of ligand required to

induce the expression of Ia. Several novel observations

by Fidler and his associates (54, 55) have indicated that

gamma IFN also may bind to receptors inside the

macrophage. These investigators showed that

liposome-encapsulated gamma IFN could initiate cellular

responses after being internalized in macrophages, even in

cells stripped of membrane binding activity by protease

treatment. Evidence for a possible nuclear location for

such internal receptors has recently been provided by

Grossberg et al. (56). In a series of ultrastructural

studies using postembedding immunocolloidal gold labeling

techniques, these workers showed that, after

internalization, gamma or beta IFN molecules accumulated

rapidly at the nuclear pores and within the nucleus of

L929 fibroblasts, primarily in association with dense

chromatin. Nuclear IFN was associated with protein

complexes of 90 kDa and 235 kDa. Beta IFN bound to









specific, saturable high affinity receptors (Kd = 1.4 X

10-10M) on the nuclear membrane (57). Comparable binding

data for gamma IFN have not yet been reported. In view of

these results, it is possible that IFN molecules

themselves may directly initiate genomic events related to

macrophage activation.

No structural information is yet available for the

binding sites for gamma IFN on the plasma membrane of

macrophages. However, on the evidence of other cell types

(51, 52, 53), the receptor is likely to be a glycoprotein

with molecular weight in the range 75-110 kDa. The

binding sites on macrophages for alpha/beta IFN also have

not been characterized in detail. The results of

competition studies using cells other than macrophages

have generally supported the idea that alpha and beta IFN

have a common receptor, while gamma IFN interacts with one

that is separate and distinct (56, 58, 59). However,

Celada et al. (49), using bone marrow culture-derived

macrophages, showed that beta, but not alpha, IFN competed

with gamma IFN for binding, suggesting that partial

overlap may occur between the two types of receptors.

Very little is known about the interactions between

triggering agents and macrophages. It is likely that

different members of this heterogeneous group function in

different ways. For example, lipid A, the active portion

of endotoxin, may initiate responses by membranous









intercalation (60), whereas maleylated proteins cause

triggering by engaging specific receptors on macrophages

(43).

Linkage of Receptor Occupancy to Effector Function

Calcium and protein kinase C. There have been

numerous reports implicating calcium in the process of

macrophage activation for tumor cell killing. Elevations

in cytosolic free calcium levels in macrophages were shown

to parallel the development of either priming in response

to gamma IFN (61) or full activation induced by lymphokine

and LPS (62). A variety of calcium antagonists and

chelators have been used to demonstrate a requirement for

intracellular, but not extracellular, calcium during

activation of macrophages by gamma IFN (or lymphokine) and

LPS (61-65); however, attempts to mimic part or all of the

activation process by using ionophores to raise

intracellular calcium levels have yielded conflicting

results. When used alone, the calcium ionophore A23187

has been reported to either fully activate (66), prime

(67), or trigger (68) macrophages for tumor cytolysis or

cytostasis, or to have no effects on activation (61, 64).

These disparate results may reflect differences in the

responding macrophage populations (e.g., bone marrow

culture-derived (67) compared to peritoneal exudate (61,

64)), concentration ranges of ionophore tested, or assay

systems used to assess cytotoxic activity. However, in









what may be an important conceptual advance, two groups

working independently have recently induced the primed

stage of activation in peritoneal exudate macrophages

exposed to low concentrations (1-10 ng/ml) of phorbol

esters in combination with A23187 (61, 64). These data

are consistent with numerous reports (reviewed in 61) that

phorbol diesters plus an ionophore for divalent cations at

low doses can cooperate synergystically to initiate a

variety of physiologic responses in cells. Such actions

have been taken as evidence for cooperation in signal

transduction between protein kinase C and elevated

intracellular levels of calcium. That protein kinase C

may be involved in macrophage activation was suggested in

a more direct way by Becton et al. (69) who showed that

priming amounts of gamma IFN caused a 2- to 5-fold

increase in the catalytic efficiency of calcium,

phospholipid-dependent protein kinase C in peritoneal

macrophages.

Requirement for protein synthesis. Whether or not

new protein synthesis is required for the completion of

the priming and triggering stages of activation is still

an unresolved issue. Most investigators have agreed that

induction of priming by gamma IFN or lymphokine takes

several hours (44, 49). This time lag between exposure to

activating agents and the expression of new function

strongly suggests that altered gene expression is a









necessary part of the process. Indeed, a variety of

qualitative changes in RNA and protein synthesis have been

documented in macrophages treated with gamma IFN or LPS

(70-72). Sharma and Piessens (73) reported that

macrophage activation by lymphokine was inhibited in a

dose-dependent fashion by either of the protein synthesis

inhibitors, pactamycin and puromycin. By contrast, in

more recent work, Hamilton et al. (72) found that

reversible inhibition of cellular protein synthesis by

cycloheximide had no effects on priming by gamma IFN when

the triggering agent LPS and the target cells were added

after removal of cycloheximide and gamma IFN. Blasi et

al. (74), in a system of activation requiring only gamma

IFN, also showed no effect of cycloheximide. However,

both the activating and triggering effects of LPS (72),

and the activating effects of alpha or beta IFN (74), were

reported to be sensitive to inhibition of protein

synthesis, suggesting to the authors that new proteins

were part of the activation program induced by these

latter agents, but were not involved in gamma IFN-mediated

priming or activation. Unfortunately, none of these

investigators addressed the possibility that cellular

signals initiated by gamma IFN in the absence of protein

synthesis could be translated into priming-associated

proteins after the removal of cycloheximide, during the









early part of the 16-24 hour coincubation of macrophages

and target cells.

Expression of Function

Tumor cell binding by macrophages. Early morphologic

studies revealed that close physical contact between

activated macrophages and tumor cell targets preceded

target cell destruction (12, 37). The capacity of

macrophages to bind tumor cells was not expressed

constitutively but could be induced by the same lymphokine

preparations which primed or activated macrophages for

tumor cell killing (75). The lymphokine which induced the

ability to bind was shown to be gamma IFN (76). Binding

competence occurred concomitantly with, and has been

proposed as a marker for, priming of macrophages for tumor

cell killing (77). Somers et al. have distinguished the

rapid, weak and nonselective binding which occurs between

all cells from that induced between gamma IFN-primed

macrophages and tumor cells. The latter was strong (>200

dynes/cell), developed over hours and occurred selectively

between primed or activated macrophages and tumor cells.

Selective binding had the characteristics of a

receptor-mediated event in that it was saturable (by

either whole cells or tumor cell membranes) and required a

trypsin-sensitive surface structure on the macrophage.

Expression of the binding site by macrophages was an

active metabolic process inhibitable by colchicine,









vinblastine, cold, and requiring transmethylation

reactions (78, 79). This last observation may be linked

to that of Bonvini et al. (80) who have recently suggested

a critical role for S-adenosylmethionine in macrophage

activation. It is of interest that the use of the phorbol

ester, PMA, was shown to bypass most of the metabolic

requirements for the expression of binding activity,

perhaps by causing the expression of a pool of cryptic

binding sites (81).

With few exceptions (e.g., 82), the evidence

currently available suggests that, under physiological

conditions (i.e., in the presence of serum proteins),

selective binding between macrophage and tumor cell is an

absolute prerequisite, but by itself is not sufficient,

for macrophage-mediated cytolysis of the target cell

(reviewed in 83).

Secretion of lytic effector molecules. Activated

macrophages secrete more than 75 defined molecules (84).

Some of these may be injurious to tumor cells (reviewed in

85). The difficulty has been in deciding which, if any,

of these potentially lytic substances has physiologic

relevance. A number of proposed cytotoxic mediators,

including arginase (86), thymidine (87), reactive oxygen

metabolites (88), and C3a (89) were effective only against

peculiarly sensitive target cell lines and are unlikely

therefore to play a central role in any general mechanism








of target cell injury by macrophages. Of particular

interest among the few cytolytic molecules or activities

that have been characterized more than superficially are

tumor necrosis factor/cachectin (TNF) and two neutral

proteases; one, discovered by Adams and his associates,

has an Mr of 40 kDa (reviewed in 85), while Reidarson et

al. (90) and Souchek et al. (91) have identified a

protease of approximate Mr 150 kd. Each of the two

cytolytic proteases was secreted only by activated,

cytolytic macrophages, and was directed principally

against neoplastic cells. These activities were

neutralized by serum, leading Adams et al. (92) to

posulate that firm, selective binding of tumor cells by

macrophages provided a diffusion-limited space where lytic

molecules could be focused onto the target cell membrane,

and from which serum protease inhibitors could be

excluded. Evidence for possible involvement of a protease

in the killing process was the observation, made initially

by Hibbs et al. (37), and confirmed by Piessens and Sharma

(93) and by Adams (94), that certain low molecular weight

serine protease inhibitors could inhibit

macrophage-mediated cytolytic activity in a dose-dependent

and irreversible manner. However, although a quantitative

assay for these lytic activities was described several

years ago (94), almost no structural or biochemical

characterization of the molecule has since been reported









and the association between cytolytic proteases and

macrophage-mediated tumor cell killing remains at the

level of phenomenology. In contrast, the recent

availability of cloned, recombinant-derived TNF (now

called TNF-a to distinguish it from the related TNF-8 or

lymphotoxin) has led to a definitive study implicating TNF

in the killing process. Urban et al. (95) reported that

tumor cells that were selected for resistance to killing

by activated macrophages were also resistant to TNF. The

converse selection of TNF-resistant variants resulted in

cells that were also resistant to lysis by macrophages.

Most importantly, it was shown that killing of tumor cells

by activated macrophages was completely inhibited by a

highly specific anti-TNF antiserum. This last result

confirms previous reports by others (96, 97) that

macrophage tumoricidal activity could be inhibited

partially by antisera raised against crude preparations of

TNF. Also consistent with the hypothesis that TNF

mediates tumor cell killing by macrophages are the

observations that LPS induces the synthesis and secretion

of TNF by macrophages (98), that gamma IFN potentiates

this LPS effect (99), and that gamma IFN causes

upregulation of TNF receptors on tumor cells (100).

Whether or not TNF is involved in all situations where

macrophages kill tumor cells (in the absence of antibody)

has not yet been determined. It is possible that a









variety of killing methods are available to the activated,

tumoricidal macrophage (which may not all depend upon

secretion of tumorolytic molecules); the particular

mechanism or combination of mechanisms utilized could then

be varied under different conditions.



Molecular Markers for Stages of Macrophage Activation

Overview

Macrophage activation stages were defined originally

by functional assays for cytotoxicity or for selective

tumor cell binding. Although these assays have been

invaluable in defining the behavior of populations of

macrophages, there is a need for reliable molecular stage

markers that could be used to examine activation at the

single cell level. Such markers would have many important

applications in research and medicine. For example, they

could be used to determine, for the first time, whether or

not all cells within a population of macrophages respond

to priming and/or triggering agents and, among those cells

that do respond, whether they traverse the stages of

activation at similar or different rates. Even more

importantly, antibodies made against stage marker proteins

could be used to assess and monitor the activation status

of macrophages in biopsy specimens of tumors or other

tissues such as immune granulomas.









Many investigators have compiled lists of metabolic

or other changes (e.g., 101) that occur in macrophages

which also happen to be activated for tumor cell killing,

but none has rigorously linked any of these changes to the

process of activation. To do so it will be necessary to

associate candidate markers with the acquisition and loss

of function in primed and activated macrophages induced by

a variety of stimuli, both in vivo and in vitro. In the

final analysis, a marker whose expression in macrophages

correlates absolutely with a stage of activation is likely

to be involved mechanistically in the activation process.

Although there is, as yet, no marker which will resolve

all of the stages of activation at the level of the

individual macrophage, there have been many reports that

purport to relate a particular molecular change or event

to at least one activation stage in populations of

macrophages. The following is a review and analysis of

the most important of these reports.

Imbalanced RNA Accumulation

The work of Varesio (102) has shown clearly that

fully activated, i.e., cytolytic, macrophages synthesize

RNA at a lower rate than other macrophages. The decrease

in overall RNA synthesis was due almost entirely to a

selective defect in the accumulation of 28S ribosomal RNA

(rRNA) (70). The other major species of rRNA, 18S, was

unaffected. This association between imbalanced rRNA









accumulation and the expression of cytolytic activity was

found in macrophages activated either in vitro by

lymphokine and LPS, or in vivo by the injection of C.

parvum. Macrophages primed by lymphokine alone had a

normal pattern of rRNA. Whether changes in RNA metabolism

are causally related to the process of activation, or

merely consequences of altered cellular metabolism, has

not been determined. However, Torres and Johnson (32)

have speculated that key cellular enzymes which normally

are induced by IFN and regulated by double-stranded RNA

(reviewed in 103), may also be influenced by

self-annealing rRNA. In this way, double-stranded regions

of rRNA may supply an endogenous signal equivalent to the

triggering action of the synthetic double-stranded RNA,

poly inosinic acid: polycytidylic acid (poly I:C) (104).

Cyclic AMP-Dependent Protein Kinase

Although cyclic AMP (cAMP) has been ascribed

regulatory roles in several macrophage functions (reviewed

in 105), no consistent link between intracellular cAMP and

macrophage activation has been established. Snider et al.

(106) found a correlation between an increase in the

levels of cAMP and activation, whereas Schultz and

coworkers (107) reported that agents that raised

intracellular cAMP also depressed the antitumor activity

of the macrophage. Other investigators have speculated

that cAMP may exert regulatory effects by its subcellular









redistribution (108) or via altered substrate molecules

(105), rather than by overall changes in concentration.

The regulatory-effects of cAMP are mediated by

cAMP-dependent protein kinases (cAMPdPK) that bind cAMP

via specific sites on the regulatory subunits of the

enzyme. Justement et al. (105) has identified 2 isozymes

of cAMPdPK (RI and RII) in the cytosol of unstimulated RAW

264.7 macrophages. Loss and reacquisition of the RI

subunit (m.w. 45.5 kDa) from the cytosol was shown to

correlate very closely with the development and loss of

cytolytic ability in both dose-response and time course

studies of activation.

Antigenic Markers

Although a great many different epitopes have been

detected on the surfaces of macrophages (reviewed in 109),

most have served to demonstrate the heterogeneity within

populations of macrophages, rather than to mark activation

stages. However, several antibodies have been described

which reacted with apparently novel antigens on fully

activated macrophages. Kaplan and Mohanakumar (110),

using an extensively absorbed rabbit antiserum raised

against the P388D1 macrophage cell line, detected a

surface antigen, AM4CSA, which was expressed by

macrophages activated in vivo by C. parvum or pyran

copolymer, but was not found on macrophages elicited by

sterile irritants. More recently Taniyama and Watanabe









(111) produced a rat monoclonal antibody with similar

specificity. ACM1 reacted with polypeptides of 45 kDa and

70 kDa on activated macrophages elicited with C. parvum or

pyran copolymer but not with resident or inflammatory

peritoneal macrophages. Two other monoclonals, MA 158.2

and 1A2.10D, described by Koestler et al. (112) and Someya

(113), respectively, also reacted more strongly with

macrophages activated in vivo than with resident

macrophages. The antigen detected by MA 158.2 was also

upregulated 3-fold when macrophages were activated by

lymphokine. However, both monoclonals lacked specificity:

MA 158.2 also reacted with inflammatory peritoneal

macrophages and 1A2.10D bound to a variety of

unstimulated macrophage cell lines. More importantly,

there is recent evidence that the antigen detected by MA

158.2 is an indicator of prior exposure of macrophages to

gamma IFN, rather than of macrophage activation per se.

Koestler et al. (114) showed, in a system free from

contaminating LPS, that macrophages primed by gamma IFN,

but not by alpha or beta IFN, expressed as much 158.2

antigen as fully cytolytic macrophages. Critical studies,

such as the one by Koestler et al., have not been reported

for any of the other antibodies described in this section.

In order to assess the specificity and future utility of

these potential stage markers, it will be necessary to









test them against macrophages in each of the stages of

activation, including the primed stage.

Proteins Detected Electropheretically

SDS polyacrylamide gel electrophoresis has been used

to analyze the more than 50 proteins secreted by

macrophages (68, 115). Grand-Perret et al. (68, 115)

related the absence or presence of particular proteins in

culture supernates to macrophage cytolytic activity.

Whole cell and plasma membrane proteins from macrophages

have also been examined electrophoretically (72, 116), and

Hamilton et al. (72) identified a series of proteins

induced by LPS under conditions which led to activation

for tumor cell killing. The development by O'Farrell

(117) of high resolution two-dimensional gel

electrophoresis has greatly facilitated the analysis of

complex mixtures of proteins. This powerful technique,

which is the basis for the research described in the

following chapters, has not previously been applied to the

analysis of macrophages in different stages of activation.

Largen and Tannenbaum (71), however, recently used 2-D

electrophoresis to resolve several hundred cellular

proteins in various populations of macrophages of unknown

activation status. No attempt was made to link any of

these proteins to tumoricidal function.









Ectoenzymes

Edelson and his colleagues, in the 1970s, developed

profiles of specific enzyme activities which distinguished

the plasma membranes of resident macrophages from those of

peritoneal macrophages stimulated in vivo (reviewed in

118). These observations were extended by Morahan et al.

(119), who sought to relate the expression of 3 of these

ectoenzymes on macrophages to cytolytic function. Of the

3 enzymes examined, 5'-nucleotidase, leucine

aminopeptidase, and alkaline phosphodiesterase, only

alkaline phosphodiesterase was modulated in a way that

appeared to correlate with activation. This enzyme

activity was 2- to 3-fold lower in pyran copolymer- and

C. parvum-activated macrophages, or in macrophages

activated by lymphokine, than it was in resident or

inflammatory macrophages. However, more recent work by

Grand-Perret et al. (68) revealed that alkaline

phosphodiesterase activity was depressed as much in primed

macrophages as it was in those that were fully activated.

Thus, ectoenzyme profiles, like other purported markers

for activation, fail to distinguish among the 3 major

stages of activation in macrophages: the unstimulated,

primed and cytolytic stages.









Summary

The preceding review of macrophage activation has

highlighted the rapid progress which has been made toward

understanding the identities and actions of the external

signals which cause macrophages to become tumoricidal. It

has also revealed how little has been learned about the

cellular processes in macrophages which connect receptor

engagement by these signal molecules to the execution of

cytolytic function. One of the principal difficulties

that has been encountered in attempting to define

activation-related subcellular events has been the extreme

heterogeneity of the macrophage populations which are

available for analysis. To overcome the problems

associated with this cellular heterogeneity, it will be

necessary to examine the process of activation at the

level of the individual macrophage. Currently, such an

approach is not practicable, as there are no reliable

molecular markers that could be used to assess the

activation status of single cells.

In the research that follows, I will describe a

systematic analysis, by two-dimensional gel

electrophoresis, of the cellular proteins of macrophages

in various stages and conditions of activation for tumor

cell killing. From these analyses, I will propose that

the expression of three proteins (or, more accurately, a

single protein plus a doublet protein), which I have









identified and designated p47b and p71/73, respectively,

accurately reflects the activation status of macrophages

regardless of the source or the nature of the activating

stimulus. I will further suggest that these potential

stage markers, which are the first proteins to be

rigorously associated with all of the stages of macrophage

activation, will have great utility in future studies of

the mechanism of macrophage activation for tumor cell

killing and, possibly, in the future, investigation of the

therapeutic efficacy of immunomodulators.
















CHAPTER II

PROTEIN CHANGES ASSOCIATED WITH STAGES OF ACTIVATION
OF BONE MARROW CULTURE-DERIVED MACROPHAGES

Introduction

Macrophages can be activated to kill tumor cells

selectively. Activation in vitro is dependent upon

exposure of macrophages to a priming agent, usually an

interferon (IFN; alpha, beta or gamma) (20, 120, 121) or

an IFN inducer (31). For most efficient expression of the

cytolytic potential of IFN-treated macrophages an

additional extracellular signal is required. This second

or triggering signal can be supplied by certain microbes

(41, 42) or microbial cell wall components, such as

bacterial lipopolysaccharide (LPS) (38, 39).

While the functional changes in populations of

macrophages treated with priming and triggering agents are

now well characterized, very little is known about the

subcellular events that underly the process of macrophage

activation for tumor cell killing. As a first step toward

understanding the molecular basis for activation, we have

used the technique of two-dimensional (2-D) gel

electrophoresis to identify protein changes that occur

after macrophages are exposed to either priming or









triggering signals, or both. Comparison of the onset of

new protein expression with the development of cytolytic

activity has allowed us to associate some of these protein

changes with activation for tumor cell killing.



Materials and Methods
Tissue Culture Media and Reagents

Methionine (meth)-free medium was prepared in our

laboratory by omitting methionine from Eagle's formula for

minimal essential medium (122). Salts for this purpose

were purchased from Fisher Scientific (Pittsburgh, PA),

amino acids from Sigma Chemical (St. Louis, MO), and

vitamins from Flow Laboratories (McLean, VA). Complete

modified Eagle's minimal essential medium was prepared

from a powdered mix (Auto-POW MEM, Flow Laboratories).

Both meth-free and regular media were supplemented with 2

mg/ml sodium bicarbonate, 2 mM glutamine (both from Flow

Laboratories), 100 U/ml injectable penicillin G potassium,

100 ug/ml injectable streptomycin (both from Pfizer, New

York, NY), and 15 mM HEPES (Sigma Chemical). Fetal bovine

serum (FBS) was obtained from Sterile Systems Inc. (Logan,

UT). All tissue culture media and sera were negative for

detectable endotoxin, as determined by assay (123) with

Limulus amebocyte lysate (LAL) (Associates of Cape Cod,

Woods Hole, MA), sensitivity 0.05 ng/ml endotoxin. Cloned

mouse gamma IFN, produced in E. coli by recombinant DNA









technology (124) and purified to a specific activity of

1.9 x 107 U/mg, was kindly provided by Genentech Inc.

(South San Francisco, CA). This preparation was negative

for endotoxin when tested at 4 x 104 U/ml, 2,000 times the

highest concentration used. The immunoglobulin fraction

of a sheep antiserum specific for mouse alpha/beta IFN was

a gift from Dr. Donna Murasko, Medical College of

Pennsylvania (Philadelphia, PA). One ml of this

preparation was capable of neutralizing 2 x 106 U of

alpha/beta IFN. An antiserum to mouse gamma IFN, produced

in rabbits, and capable of neutralizing 5 x 103 U of gamma

IFN per ml was generously provided by Dr. Howard Johnson,

University of Florida (Gainesville, FL). Bacterial LPS as

the lipid A-rich fraction II of phenol-extracted

Escherichia coli 0111:B4, was a gift from Dr. David C.

Morrison, University of Kansas School of Medicine (Kansas

City, KS).

Bone Marrow Culture-Derived Macrophages

Bone marrow was harvested from 6- to 9-week-old male

C3H/HeN mice (Charles River Laboratories, Kingston, NY)

and cultured as described (125). Briefly, cells were

flushed from tibial and femoral bone marrow cavities by

irrigation with 2 ml HEPES-buffered MEM (HMEM) expelled

through a 26 gauge needle. Cells were seeded into

bacterial grade Petri dishes (American Scientific

Products, McGaw Park, IL) at 4 x 106 cells per dish, in









HMEM supplemented with 10% v/v FBS, 5% v/v horse serum

(Flow Laboratories) and 15% v/v L-cell conditioned HMEM.

At harvest, cells were dislodged from Petri dishes by

scraping with a rubber policeman, pelleted by

centrifugation at 200xg for 10 min at 40C, and resuspended

in meth-free HMEM/10% FBS. Thirteen day cultures

contained >99% macrophages, as determined by differential

analysis of cytocentrifuge preparations stained with

Diff-Quik (Harleco, Gibbstown, NJ). Monolayers were

prepared by seeding 5 x 104 macrophages into each

flat-bottomed well of a 96-place tissue culture plate

(Costar, Cambridge, MA). The cells were incubated for at

least 2 hr at 370C in 5% CO2 in air before they were used

in cytotoxicity or radiolabeling experiments.

Assay For Macrophage-Mediated Cytolytic Activity

Killing of radiolabeled P815 mastocytoma cells by

bone marrow culture-derived (BMCD) macrophages was

measured using a 16 hr 51Cr release assay, as previously

described (39). Briefly, macrophage monolayers were

incubated either for 6 hr or for the periods indicated

with 0.2 ml of meth-free HMEM/10% FBS with or without

priming/triggering agents. Supernates were then aspirated

and monolayers were washed twice with phosphate-buffered

saline (PBS), and 2 x 104 51Cr-labeled P815 cells were

added to each well in 0.2 ml complete HMEM/10% FBS. Each

treatment was assayed in duplicate. Sixteen hr later, the









uppermost 0.1 ml of supernate was removed from each well

and assayed for radioactivity in a gamma spectrometer.

Results were expressed as percent specific 51Cr release,

as calculated by the following formula:

percent specific Cr release =

experimental release-spontaneous release
total releasable counts-spontaneous release

Spontaneous release was measured in wells where untreated
51
monolayers were incubated with Cr-labeled P815s and

medium only. Total releasable counts were quantified

after freezing and thawing Cr-labeled P815 cells in

hypotonic medium (1:1 dilution of HMEM with distilled

H20).

In these studies, only the combination of priming

agent plus triggering agent produced fully activated,

cytolytic macrophages. Gamma IFN did not by itself induce

detectable cytolytic activity, whereas LPS at triggering

concentrations (0.4-1 ng/ml) induced either no or very

slight (less than 8% specific Cr release) expression of

cytolytic activity.

Electrophoresis and Gel Processing Reagents

Acrylamide for first dimension isoelectric focusing

gels, bis-acrylamide, riboflavin, N,N,N',N'-

tetramethylethylenediamine (TEMED), ammonium persulfate,

glycine, and sodium dodecyl sulfate (SDS) were all

obtained from Bio-Rad Laboratories (Richmond, VA).









Acrylamide for second dimension SDS-PAGE gels, Nonidet

P-40 (NP-40), dithiothreitol, 2-mercaptoethanol, tris

base, tris-HCl, phenylmethylsulfonyl fluoride (PMSF),

pepstatin A, dimethyl sulfoxide (DMSO), and 2,5

diphenyloxazole (PPO) were obtained from Sigma Chemical.

Diallyltartardiamide (DATD) was produced by BDH Chemicals,

and obtained from Gallard-Schlesinger (New York, NY).

Ampholytes were from LKB Instruments (Gaithersburg, MD).

Ultra-pure urea was from Schwarz-Mann (Spring Valley, NY).

Cell Radiolabeling for 2-D Electrophoresis

Monolayers of BMCD macrophages were each incubated,

either for 6 hr or for the periods indicated, in 0.2 ml

meth-free HMEM/10% FBS (labeling medium) containing 20 uCi

L-[ 35S]meth (specific activity 1060 Ci/mmol; Amersham,

Arlington Heights, IL) with or without priming/triggering

agents. The development of cytolytic activity was assayed

in parallel in cultures that were free of 35S-meth. At

the end of the labeling period, 35S-labeled monolayers

were washed twice in ice-cold PBS and lysed directly in

the wells in 25 ul lysis solution. The lysis solution,

which was a modification of that of O'Farrell (117),

contained 9.5 M urea, 2% w/v NP-40, 100 mM dithiothreitol,

2% w/v ampholytes as Ampholine, pH 3.5-10, and the

protease inhibitors PMSF (2 x 10-4 M) and pepstatin A

(2 x 10-6 M). Lysates were immediately frozen in a dry

ice/ethanol bath and stored at -70 C. Trichloroacetic









acid-precipitable radioactivity was measured by the method

of Garrels (126). Protein concentrations in lysates were

determined by the method of Lowry (127). Samples for 2-D

analysis each contained 0.5 to 1 x 106 cpm radioactivity

and 4-6 ug protein in a volume of 10 ul.

Preparation and Analysis of 2-D Gels

Two-dimensional electrophoresis was performed

according to O'Farrell (117), essentially as modified by

Horst and Roberts (128). Briefly, samples were loaded on

first dimension tube gels (2.5 mm x 125 mm) containing

4.24% w/v acrylamide, 0.74% w/v DATD as crosslinker, 9.5 M

urea, 2% w/v NP-40 and 2.88% w/v ampholytes (2.4% pH 5-7

and 0.48% pH 3.5-10) and isoelectric focusing was

performed without prefocusing for a total of 7,000

volt-hr. Completed first-dimension gels were equilibrated

in SDS sample buffer, then stored frozen in sample buffer

at -70 0C until they were used for second dimension

SDS-PAGE. Resolving gels (160 mm wide x 280 mm long x 1.6

mm) for SDS-PAGE were 10% w/v acrylamide and 0.27% w/v

bis-acrylamide. Electrophoresis was performed at a

constant current of 11 mA until the tracking dye reached

the bottom of the slab, about 17 hr. In some gels,

nonradiolabeled molecular weight marker proteins were

included in the agarose solution used to seal IEF tube

gels to second-dimension gels, and were coelectrophoresed

with labeled cellular proteins. All gels, except those to









be stained for molecular weight calibration, were

processed for fluorography immediately after

electrophoresis, without prior fixation, as described by

Garrels (126). Briefly, gels were dehydrated by immersion

in two successive DMSO baths for a total of 40 minutes,

then impregnated with PPO (10% w/v in DMSO) for a further

90 minutes. Finally, gels were held in water overnight

before being dried and exposed to XAR-5 radiographic film

(Eastman Kodak, Rochester, NY) at -70 0C.

Several radiographic exposures, ranging from 0.5 x

106 to 5 x 106 cpm-days, were made of each gel. Each

fluorogram from treated macrophages was compared visually,

side-by-side, with a control fluorogram from medium-

treated cells. Comparisons were facilitated by the use of

a grid overlay. Each gel pattern was compared with the

control on multiple occasions, and the positions of

protein changes were recorded on a clear plastic overlay.

In fluorograms from treated cells, a change was defined as

the appearance of a spot not found in control fluorograms,

the absence of a spot found in control gels, or a marked

increase or decrease in the size or density of a spot

found in control gels. If a particular change was found

in the same relative position on fluorograms of cells

treated with different agents (as determined by molecular

weight and pH coordinates and position of the change









relative to adjacent spots) it was scored as the same

change in each treatment group. Each different change was

numbered according to its apparent molecular weight and,

using fluorograms exposed to gels for either 0.5, 1.5 or 3

x 106 cpm-days, the subject spot was semi-quantified

visually according to size and density on a scale of 0-5.

Proteins of interest were further quantified by measuring

the total optical density of the corresponding spots with

a soft laser scanning microdensitometer (SL-504-XL; Biomed

Instruments, Fullerton, CA). A series of constitutive

macrophage proteins, which were not influenced by the

stage of activation, were scanned and quantified on each

fluorogram, so that spot densities could be normalized to

allow comparisons among fluorograms of different 2-D gels.



Results

Gamma IFN and LPS Induced Numerous Protein Changes
in Macrophages

Figure 2-1 shows the proteins that were resolved by

2-D gel electrophoresis of cellular lysates from

unstimulated macrophages, and from macrophages treated

either with a priming agent alone (gamma IFN), a

triggering agent alone (LPS), or with a combination of

priming and triggering agents, to produce fully activated,

cytolytic macrophages. In each fluorogram there are at

least 2,000 well-separated spots representing proteins


























Figure 2-1. Fluorograms of macrophage cellular
proteins. Bone marrow culture-derived macrophages were
radiolabeled with S-meth and treated in one of the
following ways before lysis and 2-D gel electrophoresis: A,
medium alone; B, 20 U/ml gamma IFN alone; C, 0.5 ng/ml LPS
alone or; D, 20 U/ml gamma IFN plus 0.5 ng7ml LPS. It was
shown in parallel that only those macrophages receiving the
combination of priming and triggering agents were fully
activated for tumor cell killing. The basic end of each
gel is to the left. Proteins used for molecular weight
(kDa) calibration were myosin (205), 6-galactosidase (116),
phosphorylase B (97.4), bovine serum albumin (66),
ovalbumin (45) and carbonic anhydrase (29).

















































.- ".*" ; *



-*


*
.0



1*


*
Q


qlD


O









with isoelectric points between 4 and 8 and apparent

molecular weights greater than 22 kDa. A total of 40

different protein changes was detected when 2-D gel

patterns of macrophages representing each of the treatment

groups were analyzed and compared with gel patterns of

unstimulated macrophages (i.e., macrophages incubated in

medium alone). All of the changes were identified in each

of 9 independent experiments. The position and apparent

molecular weight (in kDa) of each of these changes are

indicated in figure 2-1. Where multiple proteins appear

at a single molecular weight, each protein is designated

by an additional letter, e.g., 47a and 47b. At least 50

other changes were detected during the course of this

work, but they were considered to be either too

inconsistent experiment to experiment, or too minor for

inclusion in final tabulations.

Comparison of these analyses, summarized in table

2-1, shows that macrophages that had been fully activated

by gamma IFN and LPS expressed all of the 26 changes seen

in macrophages treated with gamma IFN alone plus all of

the 35 changes seen in LPS-treated macrophages. An

additional protein, p32a, was induced, albeit at low

levels, only in macrophages receiving the combination of

gamma IFN and LPS. A surprising finding was the high

amount of overlap (22 shared changes) between the arrays










TABLE 2-1

Summary of the Principal Protein Changes Detected in
Macrophages Exposed to Gamma IFN and/or LPS


Macrophage Source
Protein
Stage of Bone marrowbculturea, Peritoneum, phenotype
Activation treated with elicited by 47b 71/73


nonactivated medium proteose peptone,

LPS


nonactivated LPS none (resident) +

thioglycollate +


primed IFN-y Con A, MVE-2 + -



cytolytic IFN-y + LPS P. acnes, BCG or

MSC tumor + +


a. From Chapter II.

b. Treatment for 6 hr. LPS was 1 ng/ml, gamma IFN (IFN-y) was 20
U/ml.

c. Elicitation protocols are described in legends to figure 3-2
to 3-5.









of changes induced in macrophages by such structurally

disparate molecules as gamma IFN and LPS.

The Basis for the Overlap Between Gamma IFN and LPS
Treatment Groups

Lipopolysaccharide can induce macrophages to secrete

alpha/beta IFN (129). To determine whether or not such

secretion could explain the overlap in protein changes

seen in the gamma IFN and LPS treatment groups, we

repeated the experiments described above, including a

specific sheep anti-alpha/beta IFN antiserum in some of

the incubation mixtures. As one of the controls, a rabbit

antiserum against gamma IFN was used. As is shown in

table 2-2 and figure 2-2, anti-alpha/beta IFN antiserum

either completely eliminated or partially reduced 19 of

the 22 protein changes that LPS-treated macrophages had in

common with macrophages exposed to gamma IFN. The

antiserum also prevented six additional LPS-induced

changes, involving proteins 31,35a,41,43,44b, and 44c,

from developing. In separate experiments, each of these 6

proteins was shown to be inducible by 10-100 U/ml of

either alpha IFN, beta IFN, or combinations of alpha/beta

IFN, but not by gamma IFN (figure 2-2 and data not shown).

Therefore, of the 35 protein changes induced by LPS, 25

(19 plus 6) could be attributed either partially or

completely to the actions of secreted alpha/beta IFN. The

levels of expression of the remaining proteins in the LPS






42










TABLE 2-2

Effect of Antiserum Against Alpha/Beta IFN on LPS-Induced
Protein Changes in Macrophages


Effect of antiserum
on protein change

Complete prevention




Partial prevention

Unaffected


Enhancement


Number
affected

20




5

8


2

35


Proteins affected

29b*, 31, 34* 35a, 41
43, 44a*, 44b, 44c, 45a*,
45b*, 46*, 47b*, 47c*, 57*
66c*, 83*, 89*, 91*, 117*

29a*, 63*, 64*, 66a*, 66b*

28a*, 36, 47a, 52*, 60a,
71, 73, 87*

35b, 60b


*Protein changes that were also induced by gamma IFN
(total=22).















0
N
V
* 4


0. 0


*
.~R S
I





~%%%%

S





*


/
e


4D


*


O


Iv
e.


4,'#


S


m I
9 q'


. I



*U6

*/
W


I


*



S


I


500
r-41) a o

oo a
U) 0 C 0)

C'- 0 0
1-H

0 5U

04-1 (0
ilOO EC
) r-1 r- P4 4








.r .d r. L
00 0 0)
V 4 0



V4-iO *1-4
04. 0) 0- ,-r
4-) a) b) (0
0 r. 0 4-
i-H i-H A- t o



4O) *d *
0 o C> r



o w E 0

*0 04 --H
C: (a -l U|4-)
i-l a) r0 (d









4J05-) *H)*(
(1 P (1 @ P
a) 4- 4
















r-44C10
04| 0
Cj (U (1






r.a)~ r.s-
0 0 0







0 a a) (
a) (C p Eq N
w 4 M r H





-H04C0 (d
4j 0 p to a (d
C P 0 Ei 4J





(0 00 a)0)
44 il (1 1C
(0 C (0 4 J> 04*-
U M (0 C 0 I-4


r4Ji P4i (U.co.0
a) idl m p-i
.0 -H 0 Z 441




**.4-1aa (0 a o
0 M 1 401
S0 w4 0 0
pI. 0 r A
I4 (n 0 (a 4 -
r. 4- 4- C~




0 0 $ (*)
C -1 0< -


a)r0a 4)p

0-4 ^ I N- C
4.40-i-t 0 C a)
(a4)a) M004





0
O0 () 0 0
U 0 1) r4 4-I
CN1i-4 r. 0o
44- C 1 i-lU (03
%4 -H (N 0 3 a)


0 a)
M E .-H 0


MCD-H M U) 4.J
N -i M2 P4l



CN 41a 4-I
rz4-l-4.0 LAi
a) > (L) SC -ri
P4 p >-- 0
0 0 !Z t1-1
& 0) 441 N 04-4
r. W 0 H 441
PC 1-H 4Q n 3
a) (ar V
0 >, 04 Vn
4 (d rAl f
04 3: to Q|


I









group were either enhanced or unaffected by antiserum to

alpha/beta IFN. Anti-gamma IFN antibodies had no effects.

Also, the antiserum to alpha/beta IFN had no effect on

induction of proteins by gamma IFN (data not shown).

Thus, it was concluded that there were at least two causes

of the overlap between the LPS and gamma IFN treatment

groups: (i) the autocrine effects of alpha/beta IFN

secreted by LPS-stimulated macrophages, which accounted

for most of the shared protein changes; and (ii) a

mechanism that was independent of IFN secretion.

At the conclusion of these experiments, the focus of

the work was shifted toward identifying proteins, which by

their presence or absence could mark the stages of

activation for tumor cell killing. It had become apparent

that no single protein marked a stage of activation.

Protein 32a, which was initially promising, in that it

required both gamma IFN and LPS in order to be expressed,

was ruled out when it was found to be absent from fully

cytolytic macrophages which had been activated by exposure

to LPS alone at high concentrations. We next looked for

proteins that were expressed at levels that paralleled the

expression of function, either priming or triggering for

tumor cell killing. Protein 47b, induced by gamma IFN,

and the coordinately expressed pair of LPS-induced

proteins, p71 and p73 (hereafter referred to as p71/73),

were selected for further evaluation (figure 2-3). These
















MED IFN
AILiNF Al ONF


LPS
05NG


LPS
lONG


IFN +
LPS


p47b
- b, gh n b i


p73
p71


28


80


Figure 2-3. Induction of proteins 47b and 71/73 in
macrophages exposed to priming/triggering agents. From
left to right, treatments were as follows: medium alone;
10 U/ml gamma IFN alone 0.5 ng/ml LPS alone; 10 ng/ml LPS
alone; 10 U/ml gamma IFN plus 0.5 ng/ml LPS. Each set of
2 panels corresponding to a particular treatment is a
photographic enlargement taken from individual fluorograms
similar to those shown in figure 2-1. The cytolytic
response of each teatment group, measured in parallel as
percent specific Cr release from P815 mastocytoma
targets, is shown below the corresponding pair of panels.









proteins were undetectable in unstimulated cells, strongly

expressed in activated cells, and appeared to be

relatively specific for particular inducing signals: in

fluorograms of gamma IFN-treated macrophages, p71/73 was

either undetectable or present at levels less than 3% of

those found in LPS-treated cells whereas p47b, as is

described above, was only slightly induced by LPS, and

strongly induced by exposure of macrophages to gamma IFN.

Changes in Expression of Proteins 47b and 71/73 Correlated
with Cytolytic Activity Expressed by Macrophages Exposed
to Varying Concentrations of Gamma IFN and LPS

Comparison of the appropriate panels in figure 2-4

shows that, in the presence of the priming signal, the

intensity of expression of p71/73 in macrophages was

closely related to the amount of cytolytic activity as LPS

concentration was increased (compare panels in rows 1 and

3, left column). Conversely, in the presence of a

triggering signal, p47b levels paralleled the development

of tumor cell killing as gamma IFN concentration was

increased (compare panels in rows 1 and 2, middle column).

Thus, the full development of cytolytic activity appeared

to be associated with maximal expression of both p47b and

p71/73. Absence or partial expression of either p47b or

p71/73 correlated with either noncytolytic or partially

activated macrophages, respectively. To further test the

association between protein expression and function,
























Figure 2-4. Comparison of the levels of expression of
p47b and p71/73 with the cytolytic responses in bone marrow
culture-derived macrophages. Macrophages were treated
either with varying concentrations of LPS (0-400 pg) in the
presence of a constant amount (10 U/ml) of gamma IFN (left
column); varying concentrations of gamma IFN (0-10 U/ml) in
the presence of a constant amount (400 pg/ml) of LPS
(middle column); or varying concentrations of LPS alone
(0-500 ng/ml; right column), either in the absence (solid
lines) or presence (dashed lines) of an antiserum to
alpha/beta IFN. The antiserum was used at a final
concentration sufficient to neutralize 3000 U/ml alpha/beta
IFN. Activation of macrophages for tumoricell killing (row
1) was measured as isotope release from Cr-labeled P815
mastocytoma cells as described in Materials and Methods.
Each point on the curves depicting levels of specific
protein expression (rows 2 and 3) reflects the total
optical density within the corresponding spot in 2-D
fluorograms and is given as a percentage of the maximal
densitometric value for that protein during the series of
treatments shown in this figure.























CONST. LPS-VARIABLE IFN


0


U? 60 -
z
40-

- 20-

o 0-
I-


S100-
a
w 80-

< 60

40-


0 10 30 100 300 1000
LPS CONCENTRATION
(pg/mi)


0 0.3 I 3 10 30
I FN-y CONCENTRATION
(U/ml)


LPS CONCENTRATION
(ng/ml)


100 -

80 -

60 -

40 -

20-

0-


100 -


CONST. IFN-VARIABLE LPS


LPS ALONE









macrophages were directly activated by high concentrations

of LPS alone, a process that is dependent upon the

secretion and action of alpha/beta IFN (31). As can be

seen in the right column of figure 2-4, both the shape of

the activation curve and the LPS concentration range over

which activation developed were quite different from those

experiments in which LPS was used at a low concentration

as a triggering agent (middle column). As in the latter

case, however, the intensity of expression of p47b

paralleled the cytolytic response, whereas the triggering

agent-specific p71/73 was maximally induced by

subactivating amounts of LPS (between 0.05 and 0.5 ng/ml).

As expected, the inclusion of antiserum to alpha/beta IFN

abrogated both tumor cell killing and p47b expression, but

had no obvious effects on the appearance of p71/73.

Reduced Expression of Proteins 47b and 71/73 Correlated
with the Loss of Cytolytic Activity by Macrophages

The data in table 2-3 reflect the complementary kind

of analysis to that shown in figure 2-4, i.e., the degree

to which protein expression is lost after cytolytic

activity is induced and then allowed to diminish with time

in vitro. After 3 hr of exposure to gamma IFN and LPS,

only p47b had been induced and macrophages were still

noncytolytic. As we have already shown, after 6 hr of

exposure to inducing agents, macrophages expressed both

sets of proteins and were fully cytolytic. Thereafter, as


















TABLE 2-3

Protein Expression During the Acquisition, Loss and
Reacquisition of Cytolytic Activity by Macrophages


Time (hr)b


Treatment


Densitometric
units


p47b p71/73


Cytotoxicity


(%sp. 51Cr release)


stimuli added3

stimuli removed

stimuli added


0
61
100
26
5
98


0,0
4,4
54,34
37,19
0,0
100,74


a. Determined as in figure 2-4.

b. Macrophages were labeled with 35S-meth for 3 hr before each
time point. For the evaluation of cytotoxicity,
Cr-labeled P815 mastocytoma cells were added at each time
point to cultures of macrophages treated in parallel.


c. Gamma IFN, 20 U/ml, plus LPS, 0.5 ng/ml.









the ability to kill was lost over the 24-hr period

following removal of activating agents (i.e., as

macrophages became postcytolytic), the expression of the

proteins of interest was lost concomitantly. Renewed

exposure to activating agents resulted both in the

reacquisition of cytolytic activity and the re-expression

of p47b and p71/73.

Phenotypes of Activation Stages

Proteins 47b and 71/73 can be used to construct

phenotypes that identify some of the stages that are

associated with the process of activation. These

phenotypes are presented in table 2-4.




TABLE 2-4

Protein Phenotypes for Stages of Macrophage Activation
for Tumor Cell Killing

Activation stage p47b p71/73

Unstimulated -
Primed +-
Activated + +
Postcytolytic -









Discussion

This study is the first to catalog the arrays of

protein changes induced in macrophages by agents that

prime or trigger for tumor cell killing, and to identify

specific proteins that can be used to construct phenotypes

for macrophages in the unstimulated, primed, or cytolytic

stages of activation.

These results are potentially important for several

reasons. First, we have shown that a large proportion of

the cellular protein response of macrophages to triggering

amounts of LPS (i.e., 0.4-1.0 ng/ml) is due not to direct

effects of LPS itself, but to the autocrine actions of

secreted alpha/beta IFN. We have also confirmed here by

different means the observation of Schultz and Chirigos

(31), that the activation of macrophages by high

concentrations of LPS (e.g., 100 ng/ml) is dependent upon

the secretion and effects of alpha/beta IFN. Therefore,

when evaluating either the functional or the molecular

responses of LPS-treated macrophages, even when the agent

is used at very low concentrations, it is necessary to

distinguish among effects due directly to LPS, effects due

to secreted alpha/beta IFN, and effects due to the

combined actions of LPS and IFN. Largen and Tannenbaum

(71) recently described several proteins that were

upregulated in populations of macrophages exposed to LPS.

Although the agent was used at a high concentration (1000









ng/ml) and for a prolonged period (44 hr), the authors did

not consider the potential contribution to the observed

protein changes of secreted alpha/beta IFN.

Unfortunately, because of major differences between the

overall 2-D gel patterns shown in the report by Largen and

Tannenbaum and those presented here, it is not possible to

determine whether or not any of their LPS-induced proteins

correspond to those I have shown to be induced by secreted

IFN.

My results are also important because, from among the

40 major protein changes seen in activated macrophages, I

have identified specific proteins that are expressed at

levels proportional to either the amount of priming signal

(p47b) or the amount of triggering signal (p71/73) to

which macrophages have been exposed. As is consistent

with the 2-signal model for activation, fully cytolytic

macrophages express both sets of proteins, whereas primed

or triggering agent-treated macrophages express either

p47b alone or p71/73 alone, respectively. By using

antibodies made against these marker proteins, it should

be possible, for the first time, to determine whether or

not all cells within a population develop a cellular

response to priming and triggering agents and, among those

macrophages that do respond, to determine whether they

traverse the stages of activation (priming, cytolytic, and

postcytolytic) at similar or different rates. Even more









important, such antibodies may enable us to detect and

quantify macrophage activation stages in situ. Thus, by

immunohistochemical staining of tissue sections from

tumors, it should be possible to determine whether or not

intratumoral macrophages are under the influence of

priming and/or triggering signals, and to predict their

stage of activation for tumor cell killing. Similarly, we

should be able to evaluate the effects of therapies

designed to modulate macrophage function. For example,

treatment of tumor-bearing mice with exogenous IFN or IFN

inducers would, if effective, be expected to increase the

amount of p47b in intratumoral macrophages. These marker

proteins may also have application to studies of

macrophages that have been exposed to stimuli other than

tumors, such as infectious agents or allografts. That

such approaches may be feasible has been suggested by the

results of my preliminary examinations, by 2-D gel

electrophoresis, of populations of mouse peritoneal

macrophages in various stages of activation for tumor cell

killing. In every case, the protein expression phenotype

(i.e., the presence or absence of p47b and p71/73) in

these preliminary studies has correlated with the

functional activation stage. Whether or not p47b and

p71/73 are induced in cells other than macrophages has not

been determined at an exhaustive level of analysis. If

other cell types do express these proteins, a









pan-macrophage antibody, such as the B23 monoclonal (130),

would have to be used, in conjunction with antibodies

against the phenotypic marker proteins, to distinguish

macrophages in histologic sections.

Phenotypic characterization of postcytolytic

macrophages, i.e., those that were activated by exposure

to gamma IFN and LPS and then allowed to lose cytolytic

activity over 24 hours in the absence of stimuli, is of

conceptual importance in understanding the process of

activation. On the basis of both their protein phenotypes

and their ability to respond to activating stimuli, these

postcytolytic BMCD were indistinguishable from

unstimulated macrophages. These data are in marked

disagreement with the results of Ruco and Meltzer (131),

who found, using peritoneal macrophages, that the

cytolytic stage of activation was short-lived and followed

by a postcytolytic stage that could not be reactivated by

exposure to fresh stimuli.

Finally, because the expression of p47b and p71/73 is

closely correlated with function, it follows that the

synthesis of these proteins might be causally related to

the development of macrophage activation for tumor cell

killing. This possibility is being investigated by

selecting and evaluating mutant macrophage cell lines

lacking the ability to express one or the other of the

marker proteins, as well as by assessing the effects of






56


anti-marker protein antibodies in functional assays for

activation and tumor cell killing.















CHAPTER III
PROTEIN PHENOTYPES OF MOUSE MACROPHAGES ACTIVATED
IN VIVO FOR TUMOR CELL KILLING

Introduction
Macrophages cultured from bone marrow and either

primed or fully activated in vitro for tumor cell killing

newly express a variety of cellular proteins (Chapter II).

Among them are p47b and p71/73. These proteins can be

used to identify bone marrow culture-derived (BMCD)

macrophages that are either primed for tumor cell killing

by gamma interferon (primed macrophages express P47b),

fully activated and cytolytic after concomitant exposure

to gamma interferon and bacterial lipopolysaccharide

(express both p47b and p71/73), or unstimulated (express

neither p47b nor p71/73). Bone marrow culture-derived

macrophages of the last kind can be either those that have

not been exposed to activating stimuli or that have

reverted to a quiescent state after activating stimuli

have been removed. Were such phenotypes characteristic of

macrophages activated by various stimuli in vivo, the

potential would exist for direct, intralesional evaluation

of the activation status of individual macrophages.









Given this potential, the purpose of the work

reported here was to document changes in the expression of

p47b, p71/73, and other cellular proteins by macrophages

that were exposed in vivo to various stimuli. The results

show that, just as in BMCD macrophages, specific stages of

activation can be identified in macrophages that have been

primed or activated in vivo. The results also confirm

that, of the approximately 2000 cellular proteins that can

be resolved in fluorograms of two-dimensional gels, p47b

and p71/73 appear to be most closely associated with

expression of the function of tumor cell killing.



Materials and Methods

Mice

C3H/HeN mice were obtained from the Animal Production

Area of the National Cancer Institute Frederick Cancer

Research Facility (Frederick, MD). BALB/c mice were from

Goodwin Laboratories (Plantation, FL) and C57B1/6 and

C3H/HeJ mice were purchased from The Jackson Laboratory

(Bar Harbor, ME). Mice were 7-12 weeks old when used.

Tissue Culture

Modified Eagle's minimal essential medium (HMEM) and

methionine (meth)-free HMEM were prepared and supplemented

as previously described (Chapter II). All tissue culture

media and sera contained less than 0.05 ng/ml endotoxin,









as determined by assay (123) with Limulus amebocyte lysate

(Associates of Cape Cod, Woods Hole, MA).

Reagents and Bacteria

Recombinant mouse gamma interferon (IFN), produced in

E. coli, was kindly provided by Dr. Paul Trotta of

Schering-Plough Pharmaceutical (Bloomfield, NJ). The

specific antiviral activity of this IFN preparation was

determined to be 3 x 106 U/mg protein in a plaque

reduction assay using vesicular stomatitis virus on L929

cells. Bacterial lipopolysaccharide (LPS) as the lipid

A-rich fraction II of phenol-extracted E. coli 0111:B4,

was a gift from Dr. David C. Morrison, University of

Kansas School of Medicine (Kansas City, KS). Proteose

peptone and Brewer's thioglycollate broth were purchased

from Difco Laboratories (Detroit, MI) and were prepared

according to manufacturer's instructions as 10% and 3%

solutions, respectively. Concanavalin A (Con A) was

obtained from Pharmacia Fine Chemicals (Piscataway, NJ).

Methyl vinyl ether copolymer II (MVE-2) was from Hercules

Inc. (Wilmington, DE). Bacillus Calmette-Guerin (BCG),

Phipp's Strain 1029, was originally purchased from the

Trudeau Institute (Saranac Lake, NY) and was kindly

provided by Dr. Joseph Shands of the University of

Florida. Formalin-killed Propionibacterium acnes

(Corynebacterium parvum) were obtained from Wellcome

Research Laboratories (Research Triangle Park, NC).









Peritoneal Macrophages

Peritoneal cells were harvested either from

unmanipulated mice (source of resident macrophages) or

from mice in which peritoneal exudate cells had been

elicited by intraperitoneal injections, as described in

the figure legends. Mice were killed by decapitation and

cells were harvested by lavaging each peritoneal cavity

with 10 ml of ice-cold phosphate buffered saline (PBS)

containing 2 U/ml heparin. Peritoneal cells were

resuspended in meth-free HMEM/10% fetal bovine serum

(FBS). After staining with Diff-Quik (Harleco, Gibbstown,

NJ) and differential analysis, a cell suspension

containing 2 x 105 macrophages was seeded into each

flat-bottomed well of 96 well tissue culture plates.

Following incubation for 30 min at 37C in 5% CO2 in air,

cell monolayers were washed vigorously 3 times with

PBS/10% FBS to remove nonadherent cells. Adherent

monolayers were in all cases >95% macrophages, as

determined by differential microscopic analysis of cells

stained with Giemsa.

Bone Marrow Culture-Derived Macrophages

Bone marrow was harvested from C3H/HeN mice and

cultured as described (125) for 10-14 days before use.

Monolayers were prepared by seeding 1 x 105 cells in

meth-free HMEM/10% FBS into each flat-bottomed well of 96

well tissue culture plates.









Assay for Macrophage-mediated Cytolytic Activity

Killing of radiolabeled P815 mastocytoma cells by

macrophages was measured using a 51Cr release assay, as

previously described (39). Briefly, macrophage monolayers

were incubated for 4-6 hr in meth-free HMEM/10% FBS with

or without gamma IFN and LPS. Monolayers were then washed

twice with PBS and 2 x 104 51Cr-labeled P815 cells were

added to each well. Sixteen hr later, supernates were

assayed for radioactivity in a gamma spectrometer and

results were expressed as percent specific 51Cr release.

Cell Radiolabeling and 2-D Gel Electrophoresis

Cell monolayers were each incubated for 4-6 hr in 0.2

ml meth-free HMEM/10% FBS containing 20-30 uCi 35S-meth

(specific activity 1040-1360 Ci/mmol; Amersham, Arlington

Heights, IL). The development of cytolytic activity was

assayed in parallel using macrophage cultures that were

free of 35S-meth. At the end of the labeling period,

macrophages were lysed and analyzed by 2-D gel

electrophoresis and fluorography, as previously described

(Chapter II).

Identification and Quantification of Selected Proteins

To identify proteins of interest (i.e.,

"activation-associated" proteins) in peritoneal

macrophages, the spot patterns in fluorograms from

peritoneal macrophages were matched with those in

fluorograms from BMCD macrophages which had been activated









by gamma IFN and LPS. To compensate for variations among

different isoelectric focusing (IEF) runs, lysates from

activated BMCD macrophages were included with every IEF

run. Each treatment was repeated and analyzed at least 3

times. A series of different radiographic exposures was

made of each gel to ensure that spot densities fell within

the linear response range of the radiographic film. Spots

in fluorogams were scanned with a soft laser scanning

microdensitometer (SL-504-XL; Biomed Instruments,

Fullerton, CA). Each protein of interest was quantified

by measuring the total optical density of the

corresponding fluorographic spot. To normalize the

effects of slightly varying exposure conditions among

different fluorograms, a series of 10 invariant spots was

identified and each of these spots was scanned and

quantified on every fluorogram. Data were then normalized

according to correction factors that ranged from 0.74 to

1.23.



Results

In work done previously (Chapter II), a group of 40

proteins was identified in BMCD macrophages, the

expression of which was consistently upregulated when the

cells were activated for tumor cell killing by the

combination of gamma IFN and LPS. Thirteen of these

proteins, those that were most strongly upregulated during









activation, are shown in figure 3-1. While exposure of

BMCD macrophages to gamma IFN was required for expression

of several of these proteins (e.g., p47b), some were

induced specifically by LPS (e.g., p 71/73), and others

could be induced by either agent (summarized in table

3-1). In the results that follow, expression of these 13

"activation-associated" proteins in BMCD macrophages is

compared to expression in macrophages that were obtained

directly from mice, after different kinds of stimulation

in vivo.

Nonactivated Peritoneal Macrophages

None of the resident or inflammatory macrophage

populations expressed the proteins that were gamma

IFN-specific in BMCD macrophages, proteins 46, 47b, and

47c (figures 3-2 and 3-3). Macrophages from short term

inflammatory exudates (i.e., those elicited by proteose

peptone in C3H/HeN or C57B1/6 mice or by LPS in C3H/HeN

mice) also failed to express at detectable levels the

proteins that were LPS-specific in BMCD macrophages, p35b

and p71/73, or expressed them at very low levels. By

contrast, thioglycollate-elicited C3H/HeN macrophages and

resident macrophages from C3H/HeN, BALB/c and C57B1/6 mice

synthesized moderate amounts of LPS-inducible proteins,

including p71/73. To determine whether or not endotoxin

of endogenous origin was the cause of p71/73 expression in

resident macrophages, cells from the LPS-hyporesponsive






















'."71 -



%.


I> ".'.

rI o


*


.0
0\ *
0


*
- ^*


**\
0. ^-


* 35


. 0


Figure 3-1. Fluorograms of bone marrow culture-
derived macrophages showing the presence or absence of
activation-induced proteins. Prior to lysis and 2-D
fectrophoresis, macrophages were radiolabeled with
S-meth during incubation with either A, medium alone, or
B, gamma IFN, 20 U/ml plus LPS, 0.5 ng/ml. Macrophages in
the latter group were cytolytic, i.e., fully activated,
for tumor cell killing. In panel B, activation-induced
proteins are indicated by their apparent molecular weights
(in kDa). For reference purposes, the expected positions
of p47b and p71/73 are also shown in panel A. The basic
end of each gel is to the left.





















TABLE 3-1

Signals Reponsible for Expression of Selected Proteins in
Bone Marrow Culture-Derived Macrophages (from Chapter II)


Inducing Agent


Priming dose
of IFN-y


Triggering dose
of LPS-


Either
IFN-Y or LPS


Proteins

induced


47a,63/64


46,47b,47c


35b, 71 /73


52,66a/b,66c


a. 20 U/ml

b. 1 ng/ml


c. Proteins as identified in figure 3-1.



















--+- *
e-1 N -. -i
N ,"6 ..





t, O
S(o) 4.L


0
* ~b


o(0)


i,- -' '' B^
-"
b0


37b .C "- +-


0
o

o 0


Figure 3-2. Fluorograms of resident peritoneal
macrophages. Prior to lysis and 2-D electrophoresis,
rsident peritoneal macrophages were radiolabeled with
S-meth during incubation for 6 hr with medium alone.
Macrophages were from C3H/HeN (A) or C57B1/6 mice (B).
The expected positions of p47b and p71/73 are shown in
each fluorogram. Also shown are any of the
activation-induced group of proteins shown in figure 3-1
that were increased in residents compared to the levels
found in unstimulated BMCD macrophages. The spontaneous
cytolytic activity ofleach macrophage population, measured
as percent specific Cr release from P815 targets, is
shown in the lower left corner of each panel. The number
in brackets is the cytolytic activity in the presence of
added LPS (1 ng/ml).

















* *. 0 "p .
a.-- .+,-k


". "
.
"-.. V*
7 *. ? $
*7" .' T'W


0
9


* *0


+.^ -; -. ... ..

2 .4 -e o +
S ,. a. S

*- .l O
t



47b ~


p
*0


0
*
* O


,r e '0 -
.- 0 .

-* .0 ^

'* '?


0..
* 0(4) "


C ^*" -' V- ""


.- _.*.* -
Sq *

*-
I,,. +^ +


*


2(0)


Figure 3-3. Fluorograms of inflammatory peritoneal
macrophages. Prior to lysis, radiolabeling and 2-D
electrophoresis, macrophages were elicited and harvested
according to the following schedules: A, 72 hr after
intraperitoneal (IP) injection of 1.5 ml proteose peptone
in C3H/HeN mice; B, as for A, but in C57B1/6 mice; C, 24
hr after IP injection of 1 ng LPS in C3H/HeN mice, and; D,
12 days after IP injection of 1 ml of Brewer's
thioglycollate broth in C3H/HeN mice. Proteins and
cytolytic activity are indicated as described in the
legend to figure 3-2.


* 0

0(6)


0(2)


Ii

+









C3H/HeJ strain of mice were examined. In vitro, C3H/HeJ

macrophages cannot be activated by LPS, but they are

responsive to alternative triggering agents (40).

Consistent with this, we have observed that heat-killed

Listeria monocytogenes, but not LPS, induces p71/73

synthesis in C3H/HeJ macrophages (unpublished data). As

is shown in table 3-2, the amount of p71/73 expressed in

resident macrophages from C3H/HeJ mice was comparable to

that found in C3H/HeN residents, indicating that

endogenous endotoxin alone was not responsible for

inducing p71/73 in these cells. Resident macrophages from

BALB/c and C57B1/6 mice consistently expressed these

proteins at higher levels than either C3H/HeN or C3H/HeJ

macrophages did (table 3-2).

Primed Peritoneal Macrophages

These macrophages were operationally defined as being

unable to kill, but capable of developing cytolytic

activity when they were exposed to 1 ng/ml LPS. The

patterns of expression of activation-associated proteins

in these macrophages, freshly elicited from C3H/HeN mice

by either Con A or MVE-2, resembled that seen in gamma

IFN-treated BMCD macrophages. Each of the gamma

IFN-inducible proteins was found in at least one of the

populations that had been primed in vivo, and p47b was

present in both (figure 3-4). With the exception of

barely detectable amounts of p35b in MVE-2 elicited


















TABLE 3-2

Expression of p71/73 in Resident Macrophages from
Different Mouse Strains




Mouse Strain Relative densitometric

units for p71/73


C57B1/6 74/54
BALB/c 62/98
C3H/HeN 43/18
CeH/HeJ 42/29




a. Each fluorographic spot was quantified as total optical
density (see Materials and Methods). The results are
expressed as a percentage of the total optical density
of the spots corresponding to p71/73 in the fluorogram
shown in figure 3-1, panel B (BMCD macrophages activated
by gamma IFN and LPS).





















46 40






. 46- .


2 (58)


. .


w
0 *


S6(75) .. .
6(75). '. .


Figure 3-4. Fluorograms of peritoneal macrophages
primed in vivo. Prior to lysis, labeling and 2-D
electrophoresis, primed macrophages were elicited and
harvested as follows from C3H/HeN mice: A, 24 hr after
the IP injection of 100 ug Con A, or; B, 7 days after the
IP injection of 1 mg MVE-2. Proteins and cytolytic
activity are indicated as described in the legend to
figure 3-2.


35b
*
*
S


I.









macrophages, proteins which were LPS-specific in BMCD

macrophages were not found in primed peritoneal

macrophages.

Cytolytic Peritoneal Macrophages

Each of the populations of peritoneal macrophages

that were cytolytic expressed most of the 13 proteins

which we previously have shown to be induced during the

activation of BMCD macrophages in vitro (figure 3-5).

With the exception of p63/64, which was never seen in

cytolytic macrophages from C57B1/6 mice, these proteins

were more abundant in macrophages from C57B1/6 and BALB/c

mice than they were in C3H/HeN macrophages. Eight of the

13 activation-associated proteins, i.e., proteins 46, 47a,

47b, 47c, 66a/b, and p71/73, were present in all of the 5

cytolytic populations that were examined. Of these

proteins, only p47b appeared to be expressed at levels

which corresponded with cytolytic activity. These results

were consistent through 3 different experiments. To

examine quantitatively the relation between marker protein

expression and cytolytic activity in peritoneal

macrophages, p47b and p71/73 were measured

densitometrically in fluorograms. As is shown in table

3-3, all cytolytic populations expressed p71/73 at levels

greater than 25% of those induced in BMCD macrophages by

exposure to 1 ng/ml LPS. For p47b, but not for p71/73,

the amount of marker protein expressed by macrophages



























Figure 3-5. Fluorograms of cytolytic peritoneal
macrophages, i.e., cells that had been fully activated in
vivo. Prior to radiolabeling, lysis and 2-D
electrophoresis, cytolytic macrophages were elicited as
follows: A, 14 days after IP injection of 700 ug
formalin-killed P. acnes in C3H/HeN mice; B, as for A, but
in BALB/c mice; C, 14 days after IP injection of 3 x 10
viable BCG organisms; D, 14 days after the IP injection of
10 MSC rhabdomyosarcoma cells. At the time they were
killed, each mouse in this last group had tumorous
infiltration of multiple peritoneal surfaces and a 2-3 cm
diameter spherical, sarcomatous mass attached to the hilar
surface of the spleen. In E, macrophages were elicited as
for A, but in C57B1/6 mice. Proteins and cytolytic
activity are indicated as described in the legend to figure
3-2.



















- -. 4 -
a 4


- A


S- 25(41)

"" 25(41)







0
S a


50(96)


,. '-. B





-447.
a. ,g.









.74 W7 4.*
-aj *--

** fb



D


3 4











*


a a
- a

a.- *


12 (81)
mam -a -


44(92)
*5 -*,


*


- 9

83(93)





















TABLE 3-3

Cytolytic Activity Compared to Extent of Expression
of Proteins 47b and 71/73 in Peritoneal Macrophages
Activated In Vivo


Relative Percent
Densitometric Specific
Eliciting Mouse
Agent Strain unitsa for Cr release


p71/73 p47b


P. acnes C3H/HeN 27/34 12 12
MSC tumor BALB/c 112/98 30 25
BCG C57B1/6 40/27 34 44
P. acnes BALB/c 62/55 45 50
P. acnes C57B1/6 52/39 60 83



a. Each fluorographic spot was quantified as described in
table 3-2 and expressed as a percentage of the
corresponding spot(s) in the fluorogram shown in figure
3-1, panel B.









corresponded closely with cytolytic activity. For

example, the most cytolytic population of macrophages

tested, that which was elicited in C57B1/6 mice by P.

acnes, expressed the highest levels of p47b, while

macrophages elicited in C3H/HeN mice by P. acnes were only

slightly cytolytic and synthesized barely detectable

amounts of p47b.

Protein Phenotypes for Peritoneal Macrophages in Various
Stages of Activation for Tumor Cell Killing

Using proteins 47b and 71/73, a phenotype can be

constructed for each population of peritoneal macrophages

(table 3-4). These phenotypes are the same as those I

previously reported for activation stages in BMCD

macrophages (Chapter II). Whether or not macrophages were

primed in vitro or in vivo, they consistently expressed

p47b. Similar, regardless of the source of macrophages

or the means of activating them, fully activated,

cytolytic macrophages expressed both p47b and p71/73.



Discussion

In this report, I show for the first time that simple

protein phenotypes, which utilize the presence or absence

of the cellular proteins 47b and 71/73, can be used to

identify macrophages which have been primed or fully

activated in vivo. Furthermore, I have shown that, among

different macrophage populations activated in vivo, there







76









TABLE 3-4

Protein Phenotypes for In Vivo-Derived and Bone Marrow
Culture-Derived Macrophages




Level of expression of specific proteinsa


Treatmentb 28 29a 29b 31 32a 32b 34 35a 35b 36 41. 43 44a


IFN

LPS

IFN


IFN

LPS

IFN


IFN

LPS

IFN


alone

alone

+ LPS


alone

alone

+ LPS


alone

alone

+ LPS


- u u U

u U u U u U u U

u u u u u u u U u U


44b 44c 45a 45b 46 47a 47b 47c 49 52 57 60a


U u u U U u U u -

u uc U uc u u u u U u u

u u U u u u U U u U u u



60b 63/64d 66a/b 66c 71/73 77 83 86 87 89 91 117


- U

u u

u U


- u u

U u

U u u


u u

- U

u u


a. Classified as either (-) = no different than in unstimulated
macrophages; or, (u) = upregulated; or, (U) = greatly upregulated
compared to expression in unstimulated macrophages.

b. Gamma IFN, 20 U/ml; LPS, 0.5 ng/ml.

c. Present inconsistently at 0.5 ng/ml LPS, always present at 1 ng/ml.

d. Protein doublets, e.g., 63/64, which appear to be coordinately
regulated are treated as a single entity.









is a quantitative relation between expression of p47b

(when p71/73 is coexpressed) and the level of cytolytic

activity. These results, taken together with my previous

findings for bone marrow culture-derived macrophages

activated in vitro (Chapter II), indicate that, under a

wide variety of activation conditions, protein phenotypes

of macrophages can accurately reflect activation status.

My data are consistent with a two-signal requirement

for the activation of macrophages in vivo: an initial

priming signal which induces the expression of p47b, and a

second or triggering signal which renders primed cells

cytolytic and leads to the synthesis of p71/73. For the

populations of cytolytic macrophages examined here, the

priming signal was provided almost certainly by gamma IFN

produced in the course of immune responses to bacterial or

tumoral antigens. The source in vivo of triggering

signals is less clear. Macrophages mobilized to the sites

of bacterial inflammation may come under the influence of

bacterial endotoxin or other microbial components with

triggering activity. Of particular relevance to my study,

Listeria monocytogenes, a gram-positive bacillus like P.

acnes, and muramyl dipeptide (MDP), a component of

mycobacteria (e.g., BCG), both have been shown to trigger

primed macrophages to become cytolytic (41, 132). Also, I

have preliminary evidence that exposure of macrophages to

L. monocytogenes causes the expression of p71/73 under the









same conditions which result in triggering of primed

cells. It is most unlikely that microbial products are

involved in the activation of macrophages in tumors or, as

in my study, in tumoral effusions. However, there is

increasing evidence that other endogeneous substances may

be involved in the triggering phase of macrophage

activation. Several muramyl peptides with MDP-like

activities have recently been identified in mammalian

tissues and fluids (133, 134). Also, poly I:C has been

shown to be a potent triggering agent in vitro (32),

raising the possibility that naturally-occurring duplex

RNA may serve the same function in vivo.

Neither resident nor inflammatory macrophages

constitutively expressed p47b, suggesting that these

nonactivated populations had not been exposed to gamma IFN

in the peritoneum. However, p71/73 was expressed

constitutively at moderate levels by resident macrophages

and, to a lesser degree, by thioglycollate-elicited

macrophages. These cell populations are sequestered

intraperitoneally for relatively long periods (e.g.,

macrophages elicited by thioglycollate were not harvested

until 12 days after injection), and it was conceivable

that p71/73 was induced in them by chronic exposure to

trace amounts of microbial products absorbed from the

adjacent gastrointestinal tract. If this was the case,

bacterial endotoxin clearly was not the only inducing









agent involved, as LPS-insensitive C3H/HeJ residents also

expressed p71/73. In contrast to residents, acute

inflammatory macrophages, such as those elicited by

proteose peptone, synthesized little or no p71/73 and

lacked most other activation-associated proteins and thus

were comparable to unstimulated BMCD macrophages. These

results are in general agreement with those of Largen and

Tannenbaum (71) who showed that several proteins that

could be induced in bone marrow macrophages by LPS were

present constitutively in resident macrophages, but not in

macrophages elicited by proteose peptone. The slight

amounts of p71/73 detected in most inflammatory

populations may result from resident macrophages being

"diluted" by large numbers of newly arrived p71/73-

negative inflammatory macrophages. Alternatively, the

expression of p71/73 in resident macrophages may actually

be downregulated by signals present in the inflammatory

milieu. Inflammatory macrophage populations will need to

be examined at the single cell level to distinguish

between these possibilities. It is of interest that LPS

injected intraperitoneally elicited a population of

macrophages expressing little p71/73 whereas the same

agent used in vitro strongly induced p71/73 expression.

These apparently conflicting results can be explained by

the period of several hours which elapses between

initiation of inflammation and migration of macrophages









into an inflammatory exudate (135); having stimulated the

inflammatory response, LPS is probably completely absorbed

from the peritoneal cavity before elicited macrophages

arrive from the bloodstream.

Specific antibodies against p47b and p71/73 should

provide the means to stage activation directly, either in

cultured macrophages or in those that reside in solid

tumors or inflammatory lesions, such as immune granulomas.

Currently, intralesional macrophages are accessible to

analysis only after lengthy mechanical and enzymatic

disaggregation procedures (136) that have the potential to

change the activation status of macrophages during the

process. Should cells other than macrophages also express

p47b and p71/73 in vivo, then intralesional studies will

necessitate the use of at least 2 antibodies, one to

identify the activation marker and the other to ascertain

whether or not the marker is in/on a macrophage. There

are extant at least 2 pan-macrophage monoclonal

antibodies, B23 (130) and F4/80 (137), that could be used

to identify macrophages.

In conclusion, we have shown here that activation

stages can be identified in macrophages stimulated in vivo

using the same protein phenotypes that were previously

constructed to identify activation stages in bone marrow

culture-derived macrophages. The importance of this

observation lies in its implications for future









investigations of macrophage activation in vivo at the

level of individual cells. Implicit in this statement,

there is the possibility that we will soon be able to

study how inflammatory lesions that contain activated

macrophages develop. As important, it may also soon

become possible to screen immunomodulators for efficacy in

the most relevant way possible, namely, by direct,

intralesional analysis of their effect on macrophage

activation.
















CHAPTER IV

MACROPHAGE ACTIVATION-ASSOCIATED PROTEINS: RELATION OF
PROTEIN SYNTHESIS TO PRIMING; EXPRESSION OF PROTEINS 47B AND
71/73 IN MACROPHAGES EXPOSED TO VARIOUS PRIMING AND
TRIGGERING AGENTS

Introduction

The process of macrophage activation for tumor cell

killing is closely associated with the synthesis by

macrophages of specific cellular proteins, which I have

designated p47b and p71/73 (Chapters II and III). Exposure

of bone marrow culture-derived (BMCD) macrophages to priming

amounts of gamma interferon (IFN) induces the synthesis of

p47b; a second or triggering signal supplied by bacterial

lipopolysaccharide (LPS) renders primed cells cytolytic and

leads concomitantly to the expression of p71/73. Macrophages

activated endogenously possess the same protein phenotypes as

their BMCD macrophage counterparts, i.e., in vivo-primed

macrophages express p47b, while fully activated, cytolytic

peritoneal macrophages express both p47b and p71/73 (Chapter

II).

In light of these observations, it appears that p47b and

p71/73 may have potential for use as markers for the stages

of macrophage activation. Such markers would, for the first









time, enable studies of activation to be made at the single

cell level, rather than on populations of macrophages.

However, before the use of these proteins as stage markers

can be justified in all situations of macrophage activation

for tumor cell killing, a number of unresolved questions must

be addressed. First, does the expression of p47b and p71/73

reflect macrophage activation status regardless of the

priming and triggering agents used for activation? My

previous studies (Chapter II) have focused on the effects of

the prototypical agents, gamma IFN and LPS; however,

alpha/beta IFN also can prime (40) and a large variety of

alternative triggering agents have been reported

(31,41,42,43). Second, based on the results of their studies

of cycloheximide-treated macrophages, Blasi et al. (74) and

Hamilton et al. (72) have concluded that new protein

synthesis is not necessary for either the priming or

activating effects of gamma IFN. How do these observations

relate to the synthesis of p47b during priming? Third, do

cells other than macrophages express p47b and/or p71/73?

This last question relates to the issue of whether or not a

specific marker, in addition to those for p47b and p71/73,

would be needed to distinguish macrophages from other cell

types in histologic sections of tumors or other tissues.

In the work described in this report, I have examined

each of these questions in detail, and have also made









preliminary assignments of p47b and p71/73 to cytosolic or

membranous (granular) subcellular compartments.



Materials and Methods

Mice

C3H/HeN mice were obtained from the Animal Production

Area of the National Cancer Institute-Frederick Cancer

Research Facility (Frederick, MD). C3H/HeJ mice were

purchased from the Jackson Laboratory (Bar Harbor, ME). Mice

were 7-12 weeks old when used.

Reagents and Media

All sera and tissue culture media, including modified

Eagle's minimal essential medium (HMEM) and methionine

(meth)-free HMEM were prepared as previously described

(Chapter II) and contained less than 0.05 ng/ml endotoxin, as

determined by assay (123) with Limulus amebocyte lysate

(Associates of Cape Cod, Woods Hole, MA). Recombinant-

derived mouse gamma IFN, produced in E. coli and purified to

a specific antiviral activity of 3 x 106 U/mg protein, was

generously provided by Dr. Paul Trotta of Schering-Plough

Pharmaceutical (Bloomfield, NJ). Mouse beta IFN, produced

and immunoaffinity purified as described (138), had a

specific antiviral activity of 3 x 10 U/mg protein and was

kindly provided by Dr. Judith L. Pace, University of Florida

(Gainesville, FL). Bacterial LPS as the lipid A-rich

fraction II of phenol-extracted E.coli 0111:B4 was a gift









from Dr. David C. Morrison, University of Kansas School of

Medicine (Kansas City, KS). Heat-killed Listeria

monocytogenes (HKLM) were a gift from Dr. Robert D.

Schreiber, Washington University (St. Louis, MO).

Polyinosinic acid: polycytidylic acid (poly I:C) was

purchased from Calbiochem Biochemicals (San Diego, CA). All

other reagents, unless otherwise specified, were purchased

from Sigma Chemical (St. Louis, MO).

Bone Marrow Culture-Derived Macrophages

Bone marrow was harvested and cultured as described

(125) for 10-14 days before use. Monolayers of BMCD

macrophages were prepared by seeding 6 x 104 cells in

meth-free HMEM/10% fetal bovine serum (FBS) into each

flat-bottomed well of 96-well tissue culture plates.

Assay for Macrophage-Mediated Cytolytic Activity

Killing of radiolabeled P815 mastocytoma cells by

macrophages was measured using a Cr release assay, as

previously described (39). Briefly, macrophage monolayers

were incubated for 6 hr in meth-free HMEM/10% FBS with or

without stimuli. Monolayers were then washed twice with

phosphate buffered saline (PBS) and 2 x 104 51Cr-labeled P815

cells were added to each well. Sixteen hr later, supernates

were assayed for radioactivity in a gamma spectrometer and

results were expressed as percent specific 51Cr release.









Isolation and Culture of Cells Other Than Macrophages.

Fibroblasts were explanted from tail fragments, grown to

confluency in culture flasks, and passed at least 3 times in

culture before being used in experiments. Nylon wool

nonadherent lymphocytes were prepared by the method of Julius

et al. (139). Briefly, spleens from 2 mice were

disaggregated and erythrocytes were lysed in NH4Cl lysing

solution. One hundred million spleen cells were loaded onto

each 12 ml nylon wool column. After 45 min incubation at

37C, nonadherent cells were collected in HMEM/5% FBS. By

differential microscopic analysis of cells stained with

Giemsa, this fraction contained 85-90% lymphocytes and 10-15%

neutrophils. Contaminating neutrophils were removed by

centrifugation (500xg for 30 min at 20C) through 60% Percoll

(Pharmacia Fine Chemicals, Piscataway, NJ). Interface cells,

which were >99% lymphocytes, were washed twice by

centrifugation in HMEM/10% FBS before use in experiments.

Neutrophils were harvested from the peritoneal cavities of

mice 3 hr after the intraperitoneal injection of 2 ml 0.2%

(w/v) sodium caseinate in PBS (140). Neutrophils (which

comprised 75-80% of caseinate-exudate cells) were collected

by centrifugation (500xg for 30 min at 20C) through a layer

of 70% Percoll. Pelleted cells were >99% neutrophils

according to differential microscopic analysis of cells

stained with Giemsa. Neutrophils were washed twice in

HMEM/10% FBS before being used in experiments.









Cell Radiolabeling, 2-D Gel Electrophoresis and
Quantification of Proteins

Cell monolayers were each incubated for 6 hr in 0.2 ml

meth-free HMEM/10% FBS containing 20-30 uCi 35S-meth. The

development of cytolytic activity was assayed in parallel

using macrophage cultures that were free of 35S-meth. At the

end of the labeling period, macrophages were lysed and

analyzed by 2-D electrophoresis and fluorography, as

described (Chapter II). Proteins of interest were quantified

as previously described (Chapter III). Briefly, each subject

spot in fluorograms was scanned with a soft laser scanning

densitometer (SL-504-XL; Biomed Instruments, Fullerton, CA),

and quantified by measuring the spot's total optical density.

Data were normalized by reference to a series of 10

invariant, constitutively expressed spots. Correction

factors ranged from approximately 0.75 to 1.25.

Cell Fractionation

For each fractionation experiment, 20 million BMCD

macrophages were plated in a 100 mm diameter tissue culture

dish (#25020, Corning Glass, Corning, NY) in 12 ml of

meth-free/10% FBS stimulation medium containing 100 U/ml

gamma IFN, 10 ng/ml LPS, and 50 uCi/ml of 35S-meth (specific

activity 1040-1360 Ci/mmol; Amersham, Arlington Heights, IL).

After incubation at 37 C for 10 hr, macrophages were washed

twice with PBS then scraped into a homogenization buffer that

contained 3 mM imidazole-HCl, pH 7.4, 0.25 M sucrose, 1 mM









phenylmethylsulfonyl fluoride (PMSF), 1 ug/ml pepstatin A,

and 1 ug/ml leupeptin. These and all subsequent steps were

performed at 4C. Cells were further disrupted either by 3

cycles each of 2 strokes in a Dounce tissue homogenizer

(first experiment) or by vigorous trituration with a Pasteur

pipet (second experiment.) Cellular debris and unbroken

cells were collected by centrifugation at 300xg for 10 min.

The pellet remaining after 2 washes in fractionation buffer

was the 'nuclear' fraction. The supernate was further

divided into 'granular' (pelleted) and 'cytosolic'

supernatantt) fractions by centrifugation at 110,000xg for 30

min. Each fraction was extracted in 0.1% (w/v) Nonidet P-40

(NP-40) for 30 min. Next, one half of each fraction was

adjusted to 1% NP-40 and extracted for 10 min, then

concentrated approximately 30-fold (to a final volume of 50

ul) by ultrafiltration (Centricon microconcentrator; Amicon

Corp., Danvers, MA), lyophilized, and redissolved in

O'Farrell lysis solution (117) for 2-D electrophoresis.

Fractions in 0.1% NP-40 were held at 4C and assayed for

enzyme activities within 24 hr. The plasma membrane-specific

enzyme, leucine aminopeptidase, was assayed by the method of

Morahan et al. (119); 8-D-glucuronidase was assayed according

to Hall et al. (141) and was used as a lysosomal marker;

citrate synthase activity was used to quantify mitochondria

and was assayed by the method of Shepherd and Garland (142);

and lactate dehydrogenase was used as a cytosolic marker and










was assayed by the colorimetric method of Cabaud and

Wroblewski (143).



Results

Comparison of p47b Expression and Priming by Macrophages
Exposed to Either Beta or Gamma Interferon

Bone marrow cultured-derived (BMCD) macrophages became

primed upon exposure to increasing concentrations of either

beta IFN or gamma IFN (figure 4-1, panel A). Operationally,

primed macrophages were defined as being constitutively

noncytolytic, but able to kill tumor cells after exposure to

a triggering agent. Priming, therefore, was quantified as

the cytolytic activity expressed by macrophages in the

presence of 1 ng/ml LPS.

To quantify the expression of p47b in IFN-treated

macrophages, the protein was measured densitometrically in

fluorograms of cells that had been labeled with 35S-meth,

lysed, and subjected to 2-D gel electrophoresis (figure 4-1,

panel B). It is apparent from a comparison of the panels in

figure 4-1 that, for beta as well as for gamma IFN, the

curves for expression of p47b closely approximate those for

priming. As was the case for elicitation of the primed

state, synthesis of half-maximal levels of p47b required

exposure of macrophages to a concentration of beta IFN, in

antiviral units, that was approximately 1000-fold higher than

that required for gamma IFN. The low levels of priming (10%)















mi 80 A 100 B
A -90
m 70-
S080
E 60 0
70
0 50- mU 60-
| 50-
T 40-
30 40
30 30
Ca 30
U) 20- 20*
S10- 10 A
CC O I I i I I I I I- -
S -2 -1 0 1 2 3 4 5 c -2 -1 0 1 2 3 4 5









Figure 4-1. Cytolytic activity and p47b expression
in BMCD macrophages exposed to beta or gamma IFN in the
presence of LPS. Macrophages were treated for 6 hr with
varying concentrations of either beta IFN (OD) or gamma
IFN (0) in the presence of 0.5 ng/ml LPS. In A, the
cytolytic5activity developed was measured as percent
specific Cr release as described in Materials and
Methods. In B, cultuEes treated in parallel with those in
A were labeled with S-meth, then lysed and analyzed by
2-D electrophoresis and fluorography. The expression of
p47b was quantified densitometrically in fluorograms and
expressed as a percentage of the maximal value obtained
during this set of experiments. Killing and p47b
expression in control cultures treated with LPS alone (A)
or medium (0 ) are also shown.









and p47b expression (8%) that occurred in the absence of IFN

in this experiment were attributable to the autocrine effects

of alpha/beta IFN, secreted by macrophages in response to

triggering amounts (1 ng/ml) of LPS, as I have shown

previously (Chapter II).

Although these data appeared to indicate a singularly

close association between p47b expression and priming

function, it could also be argued that the differential

induction of p47b by the two IFN was merely a reflection of a

more vigorous overall response of macrophages to gamma IFN,

compared to beta IFN. Given this possibility, it was of

interest to determine whether or not all IFN-induced proteins

were, like p47b, induced more efficiently by gamma IFN. That

this was not the case is shown clearly in figure 4-2 which is

a partial comparison of fluorograms from macrophages treated

with 100 U/ml of either beta or gamma IFN. While some

proteins, like p47b, were induced more strongly by gamma IFN

(46,47c,44a), others (41,44b,44c) were preferentially induced

by beta IFN, and some (45a,45b) were induced by either IFN

with approximately equal efficiency. Of those proteins that

were preferentially induced by gamma IFN, only p47b appeared

to be related quantitatively to the priming response of

macrophages (data not shown).








W


-A 41
de 47b




41
^iP-


\ A


mA


4a OF

44a


47c


4*
~O
i


Figure 4-2. Fluorograms of BMCD macrophages showing
that specific proteins are induced preferentially by
either beta or gamma IFN. Before lysis, 2-D
electrophoresis, and fl orography, macrophages were
labeled for 6 hr with S-meth during exposure to 100 U/ml
of either A, gamma IFN, or B, beta IFN. Designations for
specific proteins are those shown in figure 2-1.


*4 0


i-,e I









Effects of the Inhibition of Protein synthesis on IFN-induced
Priming and p47b Expression

My observation that the synthesis and expression of p47b

is intimately associated with priming may relate to the more

general question of whether or not new protein synthesis is a

necessary part of the priming process. To examine this

question, I exposed macrophages to beta or gamma IFN in the

presence of varying concentrations of the reversible protein

synthesis inhibitor, cycloheximide (CY). After 6 hr, IFN and

CY were removed by washing and target cells were added with

triggering amounts of LPS for a further 16 hr. The results

presented in figure 4-3 show that, even when protein

synthesis was reduced by 90% during the period of exposure to

IFN, priming by either beta or gamma IFN was virtually

unaffected. The data for gamma IFN-treated macrophages are

in general agreement with the results of others (72,74) who

used similar experimental protocols and concluded that new

protein synthesis during the period of exposure to gamma IFN

is not required for the subsequent expression of priming (72)

or activation (74). These previous workers did not examine

macrophages during the first few hours after the removal of

CY (along with IFN), however, to determine whether or not

intracellular signals initiated by IFN in the absence of

protein synthesis could be translated into priming-associated

proteins (i.e., p47b) upon the resumption of protein

synthesis. I, therefore, did so. During the 4 hr after CY




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EC2H504AH_FNT3QI INGEST_TIME 2012-02-20T21:55:17Z PACKAGE AA00009095_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES