Cell surface structures involved in the induction of macrophage activation for tumor cell killing

MISSING IMAGE

Material Information

Title:
Cell surface structures involved in the induction of macrophage activation for tumor cell killing
Physical Description:
vii, 140 leaves : ill. ; 29 cm.
Language:
English
Creator:
Hayes, Mark Patrick, 1959-
Publication Date:

Subjects

Subjects / Keywords:
Lymphokines -- physiology   ( mesh )
Macrophage Activation   ( mesh )
Killer Cells   ( mesh )
Antibody-Dependent Cell Cytotoxicity   ( mesh )
Pathology thesis Ph.D   ( mesh )
Dissertations, Academic -- Pathology -- UF   ( mesh )
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph.D.)--University of Florida, 1987.
Bibliography:
Bibliography: leaves 117-138.
Statement of Responsibility:
by Mark Patrick Hayes.
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 - 000955259
oclc - 18808402
notis - AER7833
System ID:
AA00009094:00001


This item is only available as the following downloads:


Full Text











CELL SURFACE STRUCTURES INVOLVED
IN THE INDUCTION OF MACROPHAGE ACTIVATION
FOR TUMOR CELL KILLING





By

MARK PATRICK HAYES


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
















ACKNOWLEDGMENTS

There are many people to whom I am indebted for moral,

advisory, or technical support during these studies. I

would like to thank my advisor, Dr. Stephen Russell, for

his guidance and for helping me to face adversity and learn

to overcome it. I would also like to thank the other

members of my committee, Drs. Alfred Esser, Lindsey Hutt-

Fletcher and Sue Moyer, for their advice and support during

my research. Many other faculty members in the Departments

of Pathology and Comparative and Experimental Pathology

were also helpful with suggestions throughout my graduate

studies. Particular thanks go to Dr. Howard Johnson and

Dr. Robert MacKay for keeping my academic interests alive.

I would like to thank Mike, Carol, Mitali, Judy, Sergio,

Cheryl, Kim, and Page for helping me to endure the endless

frustrations in the laboratory. I am grateful to Colleen

McElfresh for her extensive efforts in helping me prepare

this document. Finally, and most importantly, I would like

to thank my wife, Robin, for her constant support and

encouragement throughout all the trying times brought on by

my graduate career.
















TABLE OF CONTENTS


Page

ACKNOWLEDGMENTS . . ii

ABSTRACT . . v

CHAPTERS

I INTRODUCTION . . 1

The Mononuclear Phagocyte System: Differen-
tiation and Modulation . 1
Macrophage Activation for Tumor Cell Killing 6
Regulation of Activation for Tumor Cell Killing
by Gamma Interferon and Lipopolysaccharide:
Cell Surface Interactions . 14
Summary and Conclusion . 36

II THE INTERACTION OF MURINE GAMMA INTERFERON
WITH MACROPHAGES . 39

Introduction . . 39
Materials and Methods . 41
Results . . 46
Discussion . . 59

III INTERACTIONS OF PRIMING AND TRIGGERING STIMULI
WITH MACROPHAGES . 65

Introduction . . 65
Materials and Methods . 66
Results . . 72
Discussion . . 84

IV SOLUBILIZATION AND QUANTIFICATION OF THE
RECEPTOR FOR MURINE GAMMA INTERFERON 92

Introduction . . 92
Materials and Methods . 93
Results . . 97
Discussion . . 106


iii









Page

V SUMMARY, DISCUSSION AND FUTURE DIRECTIONS 111

REFERENCES . . 117

BIOGRAPHICAL SKETCH . . 139















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


CELL SURFACE STRUCTURES INVOLVED
IN THE INDUCTION OF MACROPHAGE ACTIVATION
FOR TUMOR CELL KILLING

By

MARK PATRICK HAYES

December, 1987


Chairman: Dr. Stephen W. Russell
Major Department: Pathology

Macrophage activation for tumor cell killing proceeds

through a series of stages induced by independent stimuli.

Gamma interferon is a lymphokine which primes macrophages

to respond to a second, triggering signal for complete

activation to the cytolytic stage. Purified, recombinant-

derived murine gamma interferon (MuIFN-gamma) was

radiolabelled with 1251 and used to determine binding

parameters on various macrophage populations. A single

class of specific, high affinity binding sites was revealed

on bone marrow culture-derived and peritoneal macrophages,

with a Kd from 4 x 10-10 to 7 x 10-9 M, depending upon the

MuIFN-gamma used for each experiment, and 2-4 x 104

sites/cell. Macrophages with altered responsiveness to








activating stimuli did not express these differences at the

MuIFN-gamma receptor level. However, priming of

macrophages with MuIFN-gamma resulted in a significant

down-regulation of receptor expression (50% decrease in

receptor number). Treatment of macrophages with proteases

revealed that the MuIFN-gamma receptor was resistant to

trypsin and chymotrypsin, but not to Pronase. Receptors

could be synthesized de novo and replaced on the cell

surface in 4 hours following removal of Pronase. These

cells also recovered the ability to respond to MuIFN-gamma

and LPS in the same time period. Trypsin affected the

ability of macrophages to respond to LPS and heat-killed

Listeria monocytogenes, and permitted the distinction

between LPS/HKLM and polyinosinic-polycytidylic acid

triggering responses (the latter was trypsin-resistant).

Trypsin treatment of macrophages revealed the dissociation

of the triggering response from the induction of proteins

by these stimuli. Further studies were directed toward

solubilization and enrichment of the MuIFN-gamma receptor.

Plasma membranes prepared from WEHI-3 cells retained

specific high affinity MuIFN-gamma binding sites.

Membranes were solubilized with octyl-p-D-glucopyranoside.

Solubilized receptor, assayed by a liposome precipitation

method, retained the properties of the membrane form, with

a slight decrease in affinity. Substantial enrichment of

receptors was achieved by chromatography on immobilized









MuIFN-gamma. These results have enhanced the understanding

of interactions of activating stimuli with macrophages and

provide the groundwork for further structural and

functional studies of the MuIFN-gamma receptor.


vii
















CHAPTER I

INTRODUCTION


The Mononuclear Phagocyte System: Differentiation
and Modulation

The cells comprising the mononuclear phagocyte system

possess a remarkable spectrum of functional capacities,

most of which are important for the variety of roles that

these cells play in host defense (reviewed in 190). The

mononuclear phagocyte system refers to the lineage of cells

that ultimately gives rise to circulating blood monocytes

and macrophages that exist in almost every tissue of the

body. The earliest recognized cell in this lineage is the

monoblast, which derives from a multipotential stem cell in

the bone marrow. Monoblasts mature in the marrow through

promonocytes to monocytes which are released into the

circulation; the circulating monocytes then migrate into

the tissues and complete their differentiation into

macrophages (189). The positive control of mononuclear

phagocyte differentiation is primarily maintained by the

colony-stimulating factors (CSF's). Those affecting

macrophage development are interleukin-3 (IL-3; also known

as multi-CSF), granulocyte-macrophage colony-stimulating

factor, and macrophage colony-stimulating factor (also











known as CSF-1) (112). The tissue macrophage is the

terminal differentiation stage of mononuclear phagocytes,

but monocytes and macrophages are subject to modulation of

their phenotype by a variety of mediators in their

environment. The functional plasticity of macrophages in

response to microenvironmental stimuli is a hallmark of the

mononuclear phagocyte system.

The functional versatility of macrophages has become

apparent over the last three decades, with the realization

that macrophages are not only phagocytic scavengers with

the capacity for exceptional directed motility, but also

act as secretary cells for a variety of important molecules

and as effector cells involved in immune responses,

antimicrobial activity, and the control of neoplasia (190).

The most distinguishing characteristic of macrophages is

their phagocytic activity, first described by Metchnikoff,

who observed their capacity for ingestion of microbial

organisms (113). Even unstimulated macrophages (if such a

population actually exists) maintain an impressive rate of

fluid phase pinocytosis, turning over the equivalent of

their cell surface area every 33 minutes (179). Macro-

phages are equally efficient at the phagocytosis of

particulate materials, both nonspecifically and via

immunologically specific recognition structures on the cell

surface (immune phagocytosis) (168). These structures

include receptors for the Fc region of immunoglobulins and











for complement components (105). The regulation of phago-

cytosis, therefore, can occur at two levels: 1) control of

nonspecific endocytic activity and 2) modulation of the

receptors involved in immune phagocytosis.

In addition to endocytosis, macrophages exhibit

several other complex functional capacities that are

instrumental for host defense. These include chemotaxis,

antigen presentation, and antimicrobial and antitumor

activities. Chemotaxis is regulated by gradients of

chemoattractant molecules that bind to receptors on the

macrophage cell surface (171). The most well-characterized

of these attractants are the N-formylated oligopeptides;

the other known attractants are the complement cleavage

product, C5a; the arachidonic acid metabolite, leukotriene

B4; and a less well-defined mediator termed lymphocyte-

derived chemotactic factor (LDCF) (172). In addition to

chemoattractants, there also exist inhibitors of macrophage

migration; a lymphocyte-derived mediator termed migration

inhibitory factor (MIF) that causes macrophages to

accumulate in its presence was first described in 1966

(37). Antigen presentation by macrophages is of primary

importance in the generation of immune responses (reviewed

in 188). The processing and presentation of antigens

requires the expression of Class II major histocompati-

bility antigens on the macrophage cell surface. This

expression is not constitutive, but is regulated by the











micro-environment of the macrophage (33). The modulation

of Class II antigen expression is primarily mediated by T

lymphocytes (14,15,158,178,195), although some regulation

of Ia expression in the mouse appears to be T cell-

independent (100).

The antimicrobial and antitumor activities of

macrophages are complex functions that normally do not

reach their full potential unless the macrophage has been

stimulated or "activated" by factors in its environment.

The concept of macrophage activation has been loosely

defined since its inception by Mackaness in the 1960's

(101,102). The definition of an activated macrophage

depended largely upon the functions) that were being

examined (30,37,123). Most of the classical studies of

murine macrophages utilized cells derived from the

peritoneal cavity; in unmanipulated mice, these are

referred to as resident macrophages (30). When mice are

injected intraperitoneally with a sterile irritant such as

thioglycollate broth, proteose peptone, or bacterial

lipopolysaccharide, macrophages derived from these cavities

exhibit numerous phenotypic differences when compared to

their resident counterparts; these cells have been termed

inflammatory (108), stimulated (68), or responsive (1)

macrophages. The heightened activities of these macro-

phages presumably results from their exposure to a largely

undefined mixture of inflammatory mediators. Macrophages











are also substantially affected by immunological mediators.

This was the phenomenon first reported by Mackaness

(101,102), who observed that acquired resistance to

microorganisms depended upon immune sensitization, but was

mediated nonspecifically by activated macrophages. It

quickly became evident that T lymphocyte interactions were

instrumental in stimulating macrophage activity (120),

particularly with respect to increased microbicidal

capacity (18,88,169). The importance of soluble T

lymphocyte products, termed lymphokines (42), in the

modulation of macrophage antimicrobial activity became

apparent (6,20,51,85,122).

Inflammatory macrophages and lymphokine-stimulated

macrophages both exhibit distinct alterations in many host

defense functions. Inflammatory macrophages have

substantially increased endocytic activity when compared to

resident macrophages (43), and also develop the capacity to

ingest complement (C3b)-coated erythrocytes (111) (C3b-

mediated phagocytosis is virtually nonexistent in resident

cells). Fc-mediated phagocytosis was shown to be enhanced

by lymphokines (193). Lymphokine supernatants were also

found to contain Ia-inducing factors) that were active

both in vivo (158) and in vitro (15,178,195). An

examination of the antimicrobial and antitumor activities

of macrophages eventually resulted in an understanding of

the stepwise nature of macrophage activation, i.e., that











resident macrophages or monocytes proceed through a

sequence of stages toward activation (30). The second step

is represented by the inflammatory macrophage, which

becomes more responsive to lymphokine-mediated activation

than resident macrophages (94,149,181). Activation of

macrophages is a highly regulated phenomenon that is

critical to the appropriate response of these cells for

host defense functions.



Macrophage Activation for Tumor Cell Killing

The antitumor activity of macrophages has been the

subject of intense investigation since the initial

observations by Hibbs et al., that macrophages derived from

toxoplasma-infected mice had the capacity to kill tumor

cells nonspecifically (67). Susceptibility of target cells

to killing by these macrophages correlated with high

saturation density of growth (no contact inhibition),

lectin agglutinability, and tumorigenicity (66).

The activation of macrophages for tumor cell killing

was shown to be mediated by lymphocyte products; a

lymphokine activity termed macrophage activating factor

(MAF) was first described by Piessens et al. (137).

Lymphokines referred to as macrophage cytotoxicity factor

(MCF) (99) and specific macrophage arming factor (SMAF)

(47) were independently described in the supernatant fluids

of appropriately stimulated lymphocytes. The latter factor











(SMAF) was unlike other described MAF's in that it

possessed specificity for the tumor cells used to generate

it. No further characterization of SMAF has appeared since

its original description. All of the macrophage activating

factors were determined to be products of T lymphocytes.

Closer examination of the activation phenomenon revealed

that MAF did not fully activate macrophages for tumoricidal

capacity under conditions that were free of endotoxin

contamination (130,199). This discovery led to the

definition of two more stages in the activation sequence

(68,108,150). Macrophages exposed to MAF-containing

lymphokines were referred to as primed, based on their

hyper-responsiveness to a second, triggering stimulus, most

commonly supplied by bacterial lipopolysaccharide (LPS).

These observations characterized a four-stage sequence of

activation: (1) the unstimulated, resident macrophage, (2)

the stimulated or inflammatory macrophage that is more

responsive to lymphokines, (3) the primed, noncytolytic

macrophage, characterized by its sensitivity to triggering

agents, and (4) the activated, cytolytic macrophage. These

stages have been characterized in vivo and in vitro

(68,108,150). The effects of lymphokine and LPS are

synergistic, i.e., responses to their combinations are far

greater than expected additive effects (68,108,130). In

most strains of mice, this synergy is extreme: macrophages

will not become tumoricidal when stimulated with











lymphokines alone unless trace amounts of LPS are present.

Increasing the amount of LPS decreases the requirement for

lymphokine, and the converse is also true (130). The

sequence of stimulation is also important. Macrophages

stimulated with lymphokines prior to addition of LPS become

cytotoxic, but if LPS is added prior to lymphokines, no

tumoricidal activity develops (134). This reinforces the

notion that priming results in a macrophage alteration that

enhances its responsiveness to the triggering agent (in

this case, LPS) prior to the addition of that agent. Other

substances that have been found to act as triggering agents

for macrophage activation include double-stranded RNA

(185), maleylated and acetylated proteins (74), heat-killed

Listeria monocytogenes organisms (165), and muramyl

dipeptide (154), as well as lymphokine components (108).

The stages of macrophage activation outlined above

have all been characterized by various functional and

molecular markers (30,73,103). Cohn established multiple

alterations that exist in the inflammatory or responsive

stage of activation: increased spreading, increases in

endocytic activity, ATP utilization, superoxide anion

production, secretary enzymes (most notably plasminogen

activator), and lysosomal hydrolases, and decreased 5'-

nucleotidase activity (30). Johnson et al. distinguished

the primed stage by the capacity to specifically bind tumor

cells and the fully activated stage by the secretion of a











putative cytolytic protease (73). More definitive bio-

chemical markers were characterized by MacKay and Russell

(103). Analysis of newly synthesized proteins by macro-

phages at various stages of activation by two-dimensional

gel electrophoresis distinguished specific protein

phenotypes induced by priming and triggering stimuli.

Specifically, a 47,000-dalton protein that appeared only in

macrophages that were primed (designated p47b), and a pair

of proteins of 71,000 and 73,000 daltons that appeared only

in macrophages that were treated with triggering agents

(designated p71,73), were consistently observed under these

respective conditions; concomitant expression of both p47b

and p71,73 was always associated with the activated state

(i.e., tumoricidal activity) (103).

Much progress has been made in the molecular charac-

terization of macrophage-activating factors) since the

initial descriptions by Nathan et al. (120) and Fowles et

al. (51). They described responses such as enhanced

spreading, motility, phagocytosis, and glucose oxidation

(120) as well as increased bacteriostasis (51). Piessens

et al. (137) were the first to describe a factor that

activated macrophages for tumor cell killing. Confusion as

to the precise nature of MAF has largely been due to the

imprecise definition of macrophage activation. Restriction

of this definition to the function of priming macrophages

for tumor cell killing allowed the eventual discovery that











at least one important MAF produced by stimulated T lympho-

cytes is gamma interferon (reviewed in 160). Early studies

of the physicochemical nature of MAF produced in stimulated

spleen cell cultures indicated a molecular weight of 55,000

daltons by gel filtration, sensitivity to protease and pH

below 4.0 or above 10.0, and adsorption to immobilized

Cibacron Blue (95). Four years later, spleen cell-derived

MAF was found to be inseparable from IFN-gamma by pH and

temperature sensitivity, immunoreactivity antiserumm to

IFN-gamma), and chromatography on Sephacryl S-200 (82), Con

A-Sepharose, polynucleotide columns, and HPLC gel perme-

ation columns (148). Biosynthesis and secretion of MAF was

coincidental with that of IFN-gamma by T-cell lines and

clones derived from them (79,140). The use of T-cell

hybridomas was instrumental in confirming that IFN-gamma is

a MAF (162). Using one such hybridoma, the laboratories of

Schreiber (164) and Russell (132) demonstrated the identity

of MAF and IFN-gamma. High resolution chromatographic

techniques (HPLC chromatofocusing and gel filtration) were

unable to distinguish the two activities (164), and

specific antisera to IFN-gamma, but not to IFN-a/p,

neutralized all of the MAF activity produced by Schreiber's

hybridoma (132). A confirmatory finding was that murine

IFN-gamma produced by recombinant DNA technology (54) was

active as a priming agent for tumoricidal activation, and

the ratio of priming to antiviral activities of the











recombinant IFN-gamma was identical to that observed with

the hybridoma-derived MAF (132,133). In both of these

studies, coincidental sensitivities to pH and temperature

were demonstrated, along with similar chromatographic

properties on polynucleotide columns and Matrex Gel Red A

(160). The identity of IFN-gamma as a MAF has been

supported by others in similar experiments (104, 167,191).

While the preponderance of evidence indicates that

IFN-gamma is the MAF that most laboratories have charac-

terized, there exist a few examples of MAF's that are not

IFN-gamma. The lymphokine originally described as macro-

phage cytotoxicity factor (MCF), derived from Con A-stimu-

lated C57BL/6 spleen cells, was reported to be unadsorbed

by poly(I)-Sepharose and thereby separated from interferon

activity (84). T-cell hybridoma-derived MAF production was

reported by Ratliff et al. (145) and Erickson et al. (45)

in which no concomitant production of IFN-gamma activity

was evident. The former hybridoma became unstable and was

lost. Therefore, no further characterization of this MAF

was possible. The latter study utilized a constitutive

producer of MAF; all other hybridomas have required

stimulation by a lectin for production. Finally, super-

natants from the thymoma EL-4, stimulated with phorbol

myristate acetate, were found to contain heterogeneous

macrophage activating activities, 50-75% of which could be

neutralized with anti-IFN-gamma antibodies, but a distinct











activity that could not be neutralized. The latter

activity had an estimated molecular weight of 23,000

daltons and was acid-stable (pH 2, 18 hr) (109). The

23,000-dalton EL-4-derived MAF has recently been identified

as B cell stimulatory factor-1 (BSF-1) (34). Another T

cell-derived lymphokine that has been shown to activate

human monocytes for tumoricidal function is human

granulocyte-macrophage colony-stimulating factor (GM-CSF)

(53). GM-CSF is also active in the mouse system (34). In

this study, it was also shown that combination of BSF-1 and

GM-CSF enhance the tumoricidal activity of murine

macrophages over that induced by BSF-1 alone (34).

While the T cell-derived factors are important for

macrophage activation, it has been well-established that

type I interferons (a and 0) are also effective for priming

macrophages (19,131). The induction of a/0 interferons by

many other substances offers an explanation for the

phenomenon of "direct" activation by many of these stimuli,

including lipopolysaccharide and other polyanions such as

polyinosinic:polycytidylic acid (166,185). These stimuli

induce a/0 interferon production by the macrophage, which

then primes in an autocrine manner, while the stimuli

themselves provide the triggering signal. The role of a/p

interferon in the regulation of macrophage actrivation is

more complex, however. If macrophages are pretreated with

a/0 interferons, they lose their responsiveness to IFN-











gamma and LPS for the development of tumoricidal activity

(129). The effects of a/0 interferons are clearly

dependent upon when macrophages are exposed to them,

relative to additional signals for activation. When the

a/3 interferons are offered simultaneously with a

triggering agent (e.g., LPS), activation occurs within 24

hours. If either a or 3 interferon is introduced as little

as 6 hours prior to addition of IFN-gamma and LPS, the

macrophage response to the latter stimuli is severely

diminished. The concentration of either IFN-a or -0 that

suppresses this response is far lower than that which is

required for priming (129).

Gamma interferon has been determined to regulate most

of the functions of macrophages previously attributed to

lymphokine regulation. It not only induces priming and

activation for tumor cell killing, but acts to maintain

cytolytic function in the presence of prostaglandin E2, a

down-regulating signal for activation (152,182). IFN-gamma

activates macrophages for enhanced microbicidal activity

(69,118,121), at least in part by enhancement of H202

production induced by other signals (121). Fc receptor

expression and immune phagocytosis are up-regulated by IFN-

gamma (48,55,136,194). Expression of major

histocompatibility antigens is also increased by IFN-gamma.

Both Class I (81,202) and Class II or Ia (11,78,81,119,177,

202,205) antigens are up-regulated; the latter directly











affects the antigen-presenting capacity of these

macrophages (13,119,205). Finally, IFN-gamma reduces the

proliferative capacity of macrophage precursors

(23,83,135,138) and induces their differentiation into

macrophages (64,135). Clearly, many aspects of macrophage

function are under regulation by gamma interferon. Most of

the alterations induced by IFN-gamma result in the overall

enhancement of the host defense potential of these cells.



Regulation of Activation for Tumor Cell Killing
by Gamma Interferon and Lipopolysaccharide:
Cell Surface Interactions

Murine gamma interferon is a polypeptide that is 132

amino acids in length (54). The interaction of polypeptide

hormones that are active at extremely low concentrations

with their target cells indicates the requirement for their

interaction with specific cell surface receptors. The

earliest suggestion of the need for interaction of

interferon with a cell surface structure was made by

Friedman in 1967 (52), before the distinctions of

interferon subtypes (a, P, gamma) were made (10). Branca

and Baglioni were the first to distinguish receptors for

Type I (a, 0) and Type II (gamma) interferons by

demonstrating competition for 125I-labelled human IFN-a

binding to lymphoblastoid cells by the former but not the

latter (188). The first direct studies of receptors for

gamma interferon were in 1982 by Anderson et al. (8,9).











Human gamma interferon, labelled with 125I-Bolton-Hunter

reagent, was used to characterize specific IFN-gamma

binding sites on human GM-258 fibroblasts. Scatchard

analysis (157) indicated the presence of 2400 binding sites

per cell with a dissociation constant (Kd) of 1.5 x 10-10M.

The binding was inhibited by unlabelled IFN-gamma and

partially inhibited by IFN-1, but not IFN-a. Neutralizing

rabbit antibodies to IFN-gamma inhibited interferon binding

to cells (8). No report since that time has found

competition for IFN-gamma binding by IFN-P. The results of

Anderson et al. (8) may have been caused by the use of an

impure preparation of naturally-derived IFN-gamma or IFN-P.

When human IFN-gamma was immunoaffinity-purified to

homogeneity and used to characterize IFN-gamma binding to

WISH amnion cells and FSII fibroblasts, no competition for

IFN-gamma was demonstrated with either a or 3 interferons

that had been similarly purified (127). In this study,

IFN-gamma bound to FSII cells with a Kd of 1 x 10-12M

(5,500 sites/cell) and to WISH cells with a Kd of 2.6 x 10-

12 M (19,500 sites/cell). The specificity of a/5 versus

gamma interferon receptors was later verified using

recombinant-derived forms of all three ligands (111).

Interaction of 125I-IFN-gamma with GM-258 fibroblasts

was further characterized as to internalization and

degradation of the ligand (9). Labelled IFN-gamma was

rapidly internalized at 370C with a t4 of 4-5 minutes and











subsequently degraded. Over eight hours, the uptake of

IFN-gamma exceeded the number of occupied receptors by five

times. As the ligand uptake was only moderately affected

by cycloheximide, it was concluded that receptors were

recycled. Degradation was inhibited by chloroquine,

suggesting that it took place in lysosomes (9). Recent

work has confirmed that murine IFN-gamma is internalized

and degraded by mouse macrophages at 370C. Degradation

occurred at a constant rate (7000 molecules/cell/hour) and

was inhibited by lysosomotropic amines. The receptor was

continuously recycled, due to the presence of a substantial

intracellular pool of receptors (29).

Early evidence for the presence of receptors for IFN-

gamma on macrophages was presented by Celada et al. (27).

IFN-gamma activity was absorbed by macrophages, even at

4cC, from a T cell hybridoma supernatant fluid. IFN-gamma

coupled to fluorescent microspheres bound to macrophages;

this binding was inhibited by an IFN-gamma-containing

hybridoma supernate. Finally, 125I-labelled recombinant-

derived murine IFN-gamma bound to bone marrow macrophages

in a specific and saturable manner with a Kd of 1.1 x

10-8M. There were an estimated 12,000 binding sites per

macrophage (27).

Accessibility to purified recombinant-derived

interferons resulted in numerous reports of IFN-gamma

receptor characterization in both human and mouse systems.










Receptors for recombinant human IFN-gamma were character-

ized on lymphoblastoid (Daudi) cells (98,111), HeLa cells

(98), embryonic fibroblast lines (111), peripheral blood

lymphocytes (111), numerous other human tumor lines

(16,186,187), and monocytes and monocytic cell lines

(26,50,128,143). Binding affinities ranged from 10-11M in

one laboratory (16,187) to 10-9M (26,111). The number of

receptors per cell varied widely, from undetectable to

20,000; one earlier study reported 50,000-70,000 sites per

cell on WISH cells (156). With the exception of isolated

human tumor cell lines, no cell type yet examined has

failed to express receptors for gamma interferon (187).

While the majority of receptor studies have been performed

with 125I-labelled interferons, a desire for greater

sensitivity and higher specific activities led Rashidbaigi

et al. to develop a new radio-ligand, IFN-gamma labelled

with 32P (143). Binding of 32P-IFN-gamma to U937

histiocytic lymphoma cells yielded comparable results to

129I-IFN-gamma binding to the same line (50). It was found

in the 32P-IFN-gamma study that IFN-gamma containing an

additional three N-terminal amino acids (Cys-Tyr-Cys) was

bound with 10-fold lower affinity than the natural IFN-

gamma molecule (143). The former IFN-gamma was produced as

a result of the predicted sequence from the cDNA, and has

often been used as a radioligand for IFN-gamma receptor

studies. The naturally-derived IFN-gamma was discovered to











lack these amino acids, and has a pyroglutamate at its

amino-terminus (147). These differences may offer one

explanation for the variations in affinity of IFN-gamma

binding reported in the literature.

Detailed studies of gamma interferon receptors in the

mouse were not reported until more recently (4,89,200).

Weitzerbin et al. reported characterization of the IFN-

gamma receptor on mouse L1210 leukemia cells, which

expressed approximately 4,000 receptors per cell with a Kd

of 5 x 10-oM (200). Preparation of 32P-labelled mouse

gamma interferon and its binding to L cells, 258C4.4 (a

cloned T cell line), and Nulli-SCC1 embryonal carcinoma

cells was demonstrated by Langer et al. (89). 32P-IFN-

gamma bound with Kd from 10-9 to 10-0oM to these cells, and

receptor numbers ranged from 5,000-15,000 per cell.

Interestingly, the induction of differentiation of Nulli-

SCC1 cells by retinoic acid resulted in a 2.5-5-fold

increase in receptor number (89). However, the induction

of differentiation of U937 and HL60 cells with phorbol

myristate acetate or dimethylsulfoxide failed to affect

IFN-gamma binding in a separate study (26).

An investigation of the binding characteristics of

IFN-gamma to mouse macrophages led Aiyer et al. to

postulate the existence of two classes of receptor on these

cells, of high (Kd = 9.1 x 10-*1M) and low (Kd = 2.7 x











10-9M) affinity (4). The affinity of the former site was

more consistent with the concentrations of IFN-gamma

required to elicit a functional response, in this case the

induction of Ia and H-2 antigens; however, even at the

proposed higher affinity, only about 5% of the receptors

were occupied at the half-maximal response (4). An

alternative explanation of these data, considered but not

favored by the authors, is that endocytosis of the

receptor-ligand complex, degradation of the ligand, and

recycling of the receptor at 37oC (the temperature at which

the binding experiments were performed) could result in the

curvilinear nature of the Scatchard plots described in this

report (201). Internalization of IFN-gamma has been

demonstrated by several laboratories (9,26,200). In one of

these studies, a curvilinear Scatchard plot indicating high

(Kd = 4.3 x 10-11M) and low (Kd = 5.9 x 10-10M) affinity

sites was derived at 370C, but not at 40C, on paraformalde-

hyde-fixed cells, or on isolated membrane preparations,

suggesting that the data at 370C were a result of

internalization of the ligand (26). Degradation of IFN-

gamma occurred rapidly at 370C in human cells (9,26), but

not in mouse L1210 cells, even though it was internalized

(200). In view of the fact that normal membrane turnover

in macrophages is quite extensive (179), it is not

surprising that internalization of ligands by these cells

affects binding parameters to such a significant degree.











The characterization of IFN-gamma receptors on macro-

phages was further complicated by data suggesting that

properties of the binding site on cells of this lineage

differed from those on nonhematopoietic cells (128).

Treatment of IFN-gamma with formic acid, which resulted in

95% inactivation of its antiviral activity, yielded a

ligand which competed for binding of 12SI-IFN-gamma to WISH

amnion cells, but not to monocytes. The acid-treated IFN-

gamma was able to induce HLA-DR antigen expression on WISH

cells, but not on monocytes; untreated IFN-gamma induced

HLA-DR antigens on both cell types, but only antiviral

activity in the WISH cells. Finally, equilibrium

saturation binding curves obtained on WISH cells yielded a

linear Scatchard plot, while on monocytes, an upward

concave dependency curve was obtained (128). While the

last observation may be explained by the complications of

internalization discussed above, the acid-treated IFN-gamma

data suggests that two distinct types of IFN-gamma

interaction may exist with monocytes in contrast to WISH

cells.

The regulation of expression of receptors for IFN-

gamma is poorly understood. Aside from the up-regulation

of receptor number upon differentiation of Nulli-SCC1 cells

(89), the only other known regulatory agent of receptor

expression is the homologous ligand, IFN-gamma. Cells pre-

treated with IFN-gamma were reduced in their capacity to











bind 125I-IFN-gamma in both mouse (200) and human (98)

cells. This effect was shown to be dose-dependent,

specific for IFN-gamma (Type I IFN's had no effect), and

reversible. The recovery of IFN-gamma binding after down-

regulation was blocked by cycloheximide and tunicamycin,

indicating the requirement for de novo synthesis of

glycoproteins (200).

The active turnover of receptors for gamma interferon

may provide an explanation for the general finding that

physiologically active concentrations of IFN-gamma are

quite low relative to the affinities of IFN-gamma binding,

such that only a few receptors would be occupied at these

concentrations. It has been suggested that there is an

inverse relationship between the number of receptors

expressed on the cell surface and the limiting

concentrations of IFN-gamma required for the induction of

HLA-DR antigens on various human tumor cell lines (16,186).

The kinetics of response are an important factor. A

minimal time of exposure to the ligand is apparently

required, and the length of exposure necessary to elicit a

response is inversely proportional to the concentrations of

the ligand and receptor (16). The hypothesis that has been

advanced is that the intracellular signal accumulates with

repeated cycles of receptor-ligand internalization (16,29).

It has been suggested that the number of cycles of receptor

occupancy required for induction of functions depends upon











the function, e.g., enhanced H202 secretion requires only

one round of receptor occupancy, while activation for tumor

cell killing requires three to four rounds (29).

The nature of the intracellular signals) by which

IFN-gamma induces the pleiotropic responses that it

mediates is still unclear. However, some clues to the

potential coupling events between binding of IFN-gamma and

its ultimate effects have been revealed in pharmacological

studies of IFN-gamma action. Using the down-regulation of

transferring receptors by IFN-gamma as a model, Weiel et al.

(198) reported that the calcium ionophore A23187 and

phorbol myristate acetate (PMA) each independently mimicked

this effect of IFN-gamma. Phorbol esters known to activate

protein kinase C were all similarly active in this system.

Furthermore, suboptimal concentrations of PMA and A23187

cooperatively reduced transferring receptor expression in

macrophages. The results implied potential roles of

elevated intracellular Ca*+ and stimulation of protein

kinase C in IFN-gamma action (198). The converse observa-

tion, stimulation of protein kinase C activity by treatment

of macrophages with IFN-gamma, reinforced this hypothesis

(60). Using bone marrow culture-derived macrophages,

Johnson and Torres demonstrated directly that calcium

ionophore A23187 could prime macrophages for tumor cell

killing. PMA, diacylglycerol, arachidonic acid,

leukotriene B4, and platelet-activating factor were all











ineffective in this system (75). Somers et al.

subsequently showed that PMA and A23187, in combination,

were able to prime peritoneal macrophages for tumoricidal

capacity. In addition, an intracellular Ca*+ chelator

(Quin-2/AM), but not an extracellular Ca*+ chelator (EGTA),

blocked priming of macrophages by IFN-gamma (173). These

results were confirmed and extended by Celada and

Schreiber; antagonists of calcium-binding proteins (of

which protein kinase C is one) were found to block priming

of macrophages by IFN-gamma (28). The A/J mouse provided a

model for unresponsiveness to IFN-gamma for tumoricidal

activation. Macrophages from these animals exhibit normal

binding of IFN-gamma, and can be primed with A23187 and

PMA, but do not show increases in protein kinase C activity

or efflux of intracellular calcium in response to IFN-gamma

(61). All of these results, taken together, indicate the

potential importance of protein kinase C activity and

particularly intracellular calcium levels in the regulation

of IFN-gamma-mediated priming of macrophages for tumor cell

killing. It is not yet known whether the IFN-gamma

receptor is coupled to the phospholipase-dependent cleavage

of phosphatidylinositol-4,5-bisphosphate into diacylycerol

(which activates protein kinase C) and inositol-1,4,5-

triphosphate (which increases intracellular Ca+*

concentrations).











Complete understanding of the mechanisms involved in

transduction of the IFN-gamma signal will require the

biochemical characterization of the receptor and any

associated cell surface molecules. The first attempts at

biochemical characterization of the receptor for IFN-gamma

involved the use of crosslinking techniques. Crosslinking

of 125I-labelled human IFN-gamma to WISH cells using either

disuccinimidyl suberate (DSS) or ethylene glycol bis

(succinimidyl succinate) yielded a labelled complex that

migrated with an apparent molecular mass of 105,000 5,000

daltons when analyzed by sodium dodecylsulfate

polyacrylamide gel electrophoresis (SDS-PAGE). Formation

of the complex was inhibited by preincubation with

unlabelled IFN-gamma, but not with unlabelled IFN-3 (156).

Human IFN-gamma that had been acylated with N-succinimidyl-

6 (4'-azido-2'-nitrophenylamino) hexanoate, a photoreactive

crosslinker, formed a complex upon binding to GM-258

fibroblasts and photoactivation that migrated with an Mr of

230,000 7,000 daltons (7). Several other reports using

crosslinking techniques indicated the formation of

complexes of varied molecular weights. Using DSS, Littman

et al. reported the formation of a complex of Mr = 125,000

20,000 from Daudi and HeLa cells (98). Celada et al.,

crosslinking 125I-IFN-gamma to U937 cells or membranes,

demonstrated a diffusely migrating complex of 84,000-

107,000 daltons (26). Murine IFN-gamma crosslinked to











L1210 cells yielded a complex of Mr = 110,000 5,000

daltons (200). Using 32P-IFN-gamma, Langer et al. reported

the formation of a complex of Mr = 90,000-97,000 daltons

from mouse cells (89) and 117,000 17,000 daltons from

human cells (144). The former estimate was later changed

to 110,000 15,000 daltons for the murine receptor complex

(106). A heterogeneous pattern of complexes was reported

by Ucer et al. (187) upon crosslinking of IFN-gamma to

various human tumor lines. Three distinct bands of 70 5,

92 5, and 150 10 kilodaltons were seen in all cells

tested. In all of these studies, formation of the

complexes was inhibited by prior incubation of the cells or

membranes with unlabelled IFN-gamma. The complexes formed

always migrated on SDS-PAGE to form a band of a diffuse

nature.

Direct biochemical analysis of the receptor has so far

been limited to the human system. Polyclonal antibodies to

the human IFN-gamma receptor were prepared by immunizing

rabbits with a receptor-rich cell line and extensively

adsorbing the sera with a receptor-negative cell line. The

resultant antibodies immunoprecipitated internally labelled

molecules with a Mr of 75,000 to 90,000 by SDS-PAGE (87).

The human IFN-gamma receptor has been purified by affinity

chromatography on immobilized IFN-gamma columns (2,124,

125). Novick et al. reported the recovery of three

molecular species, upon elution of Triton X-100 extracts











from an immobilized IFN-gamma column; a major band of Mr =

95,000 and two minor bands of Mr = 60,000 and 80,000 were

revealed by SDS-PAGE under reducing conditions.

Nonreducing conditions yielded a major band of Mr > 200,000

(124,125). Polyclonal antibodies produced in mice by

immunization with the 95,000-dalton protein possessed

neutralizing activity (antiviral) specific for IFN-gamma,

and immunoprecipitated a 125I-IFN-gamma-receptor cross-

linked complex of Mr = 120,000 (125). Aguet and Merlin (2)

reported purification of the human IFN-gamma receptor from

Raji lymphoma cells by sequential affinity chromatography

on an IFN-gamma column and a column to which a monoclonal

antibody prepared to enriched receptor was coupled. The

final eluate consisted of two molecular species, of Mr =

90,000 and 50,000. Both of these proteins bound IFN-gamma

on nitrocellulose blots. The two proteins migrated to the

same extent under both reducing and nonreducing conditions.

The relationship between the two is unclear. The 50,000-

dalton protein may represent a degradation product of the

90,000-dalton molecule, or the two species may be

indicative of receptor subunits. The definitive structure

of the IFN-gamma receptor awaits further characterization,

particularly with regard to putative subunits and/or

associated components that may be important for

transduction of the IFN-gamma signal. It will be of

particular interest to see if there are structural











differences between receptors on macrophages and non-

hematopoietic cells, as hypothesized by Orchansky et al.

(125,128). This hypothesis was based primarily on

differences in the ability of acid-treated IFN-gamma to

induce HLA-DR antigens and antiviral activity on WISH cells

and monocytes (128). Intact IFN-gamma appeared to be

required for antiviral activity on WISH cells and induction

of HLA-DR on monocytes. It would be of interest to

determine if other, more complex functions such as

microbicidal or tumoricidal activity by macrophages were

also dependent upon intact IFN-gamma. Acid-treated IFN-

gamma was able to induce HLA-DR on WISH cells, indicating

that both cell type and function induced were

distinguishable by this procedure. One possibility that

has precedent is that the receptor is the same on all cell

types, but contains distinct cytoplasmic domains that may

be associated with different functional signals, as with

the Fc receptor (146). The availability of antibodies to

the receptor (2,125) should prove invaluable in delineating

important structural characteristics required for IFN-gamma

action.

The absolute requirement of IFN-gamma receptor

engagement for macrophage activation and antiviral activity

has been called into question by two different sets of

experiments. First, encapsulation of IFN-gamma into

liposomes created a vehicle for IFN-gamma action that











appeared to bypass interaction with the receptor, since it

was refractory to neutralization by antibodies to IFN-gamma

and resulted in the abrogation of species (human-mouse)

specificity of the ligand (49). Although questions of

liposome leakage made these data controversial (44), the

ability to bypass the need for extracellular receptor

interaction was reinforced by transformation experiments,

in which mouse cells transformed with a human cDNA encoding

a non-secreted form of IFN-gamma acquired permanent

antiviral activity (155). These cells were transfected

with a cDNA that lacked the signal sequence required for

secretion of the ligand; these cells accumulated the IFN-

gamma intracellularly, where it could not interact with the

cell surface receptor. These experiments also suggested

that species specificity of IFN-gamma is maintained at the

receptor level.

The nature of the interaction of triggering signals

for tumoricidal activation of macrophages with these cells

is much less clear, with the possible exception of the

demonstration of receptors for maleylated proteins that

appear to mediate the secretion of neutral proteases by

macrophages (74). The most commonly utilized and well-

studied triggering agent for macrophage activation is

bacterial lipopolysaccharide (LPS) (68,108,130,150). The

existence of a specific "LPS receptor" that might mediate

the triggering signal has been the subject of controversy











(reviewed in 117). Springer et al. (176) were the first to

describe the presence of an LPS receptor on erythrocytes;

it was characterized by virtue of its ability, once

isolated, to displace 32P-labelled LPS from erythrocytes.

The receptor was isolated and purified to apparent

homogeneity; it was identified as a 256,000-dalton

lipoglycoprotein rich in N-acetyl-neuraminic acid,

galactose, and hexosamines (175). Attempts by this group

to characterize a similar receptor on leukocytes and

platelets failed. Instead, LPS-binding activity from these

cells was found to be of low specificity (binding other

bacterial antigens) and composed entirely of lipid (174).

The activity of LPS as a triggering agent, as well as

for most other functions, is entirely attributable to the

lipid A component (41). The availability of inbred strains

of mice that are hyporesponsive to the numerous effects of

LPS and lipid A provided a powerful tool for dissecting the

interactions of LPS and macrophages that lead to a response

(110). Taking advantage of such a system, Couthino et al.

prepared an antiserum reactive to the "lipid A-specific

triggering receptor" on B cells by immunizing rabbits with

C3H/Tif (LPS high responder) cells and extensively absorb-

ing the resulting serum on C3H-HeJ (LPS low responder)

cells. Lipopolysaccharide and lipid A competed for the

binding of this antiserum to C3H/TifB cells, and the anti-











body was mitogenic for these cells (32). Unfortunately,

these studies have not been reproducible (197).

Several attempts to study the binding of LPS or LPS

components directly to cells have been made in recent

years. Davies et al. described the binding of 32P- and

14C-labelled LPS to various cell types and demonstrated an

unusual cyclical absorption-desorption pattern of binding

with a periodicity of around 60 minutes (38). A specific

LPS receptor on rabbit peritoneal macrophages was

characterized using 3H-LPS as a ligand. In this study

3H-LPS was found to bind reversibly and specifically to

both resident and elicited peritoneal macrophages, but not

to lung macrophages (to which LPS bound irreversibly and

nonspecifically). It was determined that the macrophages

bound 3 x 104 LPS molecules per cell, and ligand

specificity was assessed by lack of competition with LPS

from a distinct bacterial species. The binding was

apparently mediated through the polysaccharide component of

the LPS, and it was not inhibited by the lipid A component

(58). More recent studies of this binding phenomenon

indicated a requirement for the presence of serum for

binding to occur, and implicated a potential role for the

complement component C3. Specific binding of 3H-LPS was

demonstrated for both human monocytes and mouse macrophages

in this system (56). As the majority of effects of LPS are

mediated via the lipid A region, the existence of a











polysaccharide-specific receptor may be of minimal

functional importance. However, it has been shown that the

isolated polysaccharide moiety of LPS from Bordetella

pertussis can activate B cells in a macrophage-dependent

system, and that this apparently occurs via the stimulation

of IL-1 secretion from the macrophages (57). The

interactions of 3H-LPS and 5'Cr-labelled lipid A with human

monocytes have also been studied. 3H-LPS bound rapidly (ts

< 5 minutes) and reversibly to monocytes, followed by a

gradual loss in binding over time, if the assay was

terminated using trichloroacetic acid. However, if the

assay was terminated by cold filtration, a more gradual,

irreversible binding was noted. Use of 51Cr-lipid A as a

ligand resulted in only the former type of binding. If

monocytes were pretreated with LPS, the number of binding

sites (but not the affinity) diminished (92). When

membranes were prepared from monocytes, binding of 3H-LPS

and e5Cr-lipid A was less specific and of lower affinity

than to whole cells, leading the authors to conclude that

some assembly of structures in the lipid bilayer of the

intact cell membrane could occur that were required for the

efficient high affinity specific binding of LPS and lipid

A. The speculation was that this assembly could not occur

after membranes were prepared from the cells (91). Other

studies with radiolabelled LPS have yielded results that

were not indicative of a specific receptor. Warner et al.











demonstrated binding of 125I-LPS to human monocytes that

was neither saturable with time nor LPS concentration, and

that was inhibitable by the O-polysaccharide but not by

lipid A or polymyxin B (an antibiotic that binds to lipid

A) (196). The preponderance of evidence related to binding

of LPS to cells has indicated that binding is probably

necessary, but not sufficient, for the induction of its

effects (117). One of the overriding complications of

studying the interaction of LPS with cell membranes is the

high degree of nonspecific interactions associated with

LPS. LPS interacts strongly with a number of membrane

components such as glycoproteins, glycosphingolipids, and

phospholipids (115,117). Recently, the potential

importance of gangliosides in LPS responses has been

reported (31,153). The interaction of LPS with monocyte

and macrophages has been shown to result in an increase in

plasma membrane microviscosity and a decrease in the order

of the lipid bilayer (90). There is a transient

perturbation in the membranes of primed (therefore

hyperresponsive to LPS) but not elicited macrophages upon

exposure to LPS (46), suggesting a role for this

perturbation in the triggering response for tumor cell

killing. The tendency of LPS to interact with plasma

membrane components in an avid manner may be of primary

importance to the ultimate generation of the signals

leading to the numerous responses of cells to this











molecule. Morrison and Rudbach (117) have proposed that

both intercalation of LPS in the lipid bilayer and

interaction with high affinity receptors are required for

functional responses to LPS. In this hypothesis, the lipid

bilayer interaction would precede the receptor binding.

Consistent with this model is that LPS interactions with

membranes and responses to LPS are time- and temperature-

dependent, and they are dependent upon the phospholipid

composition of the membrane bilayers. Finally, these

authors maintain that LPS-induced membrane perturbations

may be sufficient for some functional responses that

require a second signal (e.g., Con A or PMA), but not for

other responses (117). Although these are a variety of

membrane components with which LPS may interact strongly

enough to induce such perturbations, there is good evidence

for the role of gangliosides in the response of macrophages

to LPS, based on their ability to inhibit LPS-induced

functions in macrophages (31,153).

One attractive hypothesis for the interaction of LPS

with the plasma membrane and the coupling of this

interaction to responses is that LPS inserts itself into

the bilayer and subsequently associates specifically with

some membrane components) that is(are) responsible for the

genesis of a transmembrane signal. Intriguing evidence for

this hypothesis was reported by Kilpatrick-Smith et al.,

who performed photobleaching experiments using LPS











conjugated to fluorescein isothiocyanate (80). These

studies allowed the determination of the lateral mobility

of bilayer-associated LPS. It was observed that there were

two subpopulations of LPS bound to the cell, one of rapid

mobility (D = 10--'10-e cm2/s) and one of relative

immobility (D < 10-12 cm2/s) (D = lateral diffusion

coefficient). These results suggested to the authors that

at least some LPS, after insertion into the plasma

membrane, became associated with an anchored membrane

component (80).

The interaction of LPS with the cell surface that

leads to a stimulus and response is still poorly defined.

Recent evidence suggests that LPS may induce signal

transduction via the hydrolysis of phosphatidylinositol-

4,5-bisphosphate (PIP2), which would lead to increased

intracellular calcium levels (mediated by inositol-1,4,5-

triphosphate) and activation of protein kinase C (mediated

by diacylglycerol). PIP2 hydrolysis occurred with levels

of LPS that were physiological relevant to macrophage

activation (139). There is also evidence to implicate a

role for guanine nucleotide regulatory proteins in the

mediation of LPS responses by B cells and macrophages (72).

Earlier studies have suggested a possible role for

activation of cell-associated serine proteases in

generation of LPS-induced responses. Several laboratories

reported mimicry of LPS stimulation of B cells by trypsin











and inhibition of LPS stimulation by serine protease

inhibitors such as soybean trypsin inhibitor, N-a-tosyl-L-

lysyl-L-chloromethane (TLCK) and di-isopropyl

phosphofluoridate (DFP) (12,62,63,76,114). While endotoxin

contamination in trypsin preparations calls the former

observation into question, the latter results suggested

that mediation of LPS responses (by B cells) occurred via

the early action of a cellular serine protease. In support

of this hypothesis is the fact that LPS activates at least

two soluble serine proteases, factor XII of the intrinsic

coagulation system and the complement component Cl (115).

Characterization of a specific receptor for LPS that

is responsible for initiating a response should help to

clarify the means by which LPS induces its many effects. A

series of experiments reported by Wright and Jong (204)

implicate the family of adhesion-promoting receptors as LPS

receptors. Specifically, antibodies to the 95,000-dalton B

chain common to CR3 (receptor for the complement cleavage

product C3bi), lymphocyte function-associated antigen (LFA-

1), and p150,95 (a dimeric protein of unknown function)

removed the capacity of human macrophages to bind LPS (as

E. coli or LPS-coated erythrocytes). This binding was

time- and temperature-dependent, and mediated by lipid A.

The authors speculated that the receptor recognized a

common structure, such as a sugar phosphate, on the lipid A

portion of the LPS (204).











Summary and Conclusion

The important role of macrophages in host defense has

been well documented. These cells are unique in the

versatility of their functional activities, ranging from

the phagocytosis of a variety of particles, chemotaxis, and

participation in the immune response, to their microbicidal

and tumoricidal functions. Virtually all of the activities

of macrophages are induced or regulated by environmental

stimuli. The concept of macrophage activation has been the

subject of intense investigation since its original

description by Mackaness (101,102). Although the

phenomenon of macrophage activation for tumor cell killing

has been well described (1), the sub-cellular and molecular

events involved in the induction of activation are poorly

understood. Characterization of these molecular

interactions is critical to the ultimate understanding of

how macrophage tumoricidal activities are regulated.

Significant progress in the delineation of the

molecular events involved in macrophage activation was made

possible by the identification of the stimuli required for

the induction of the priming and triggering stages of

activation. Of central importance was the cloning and

purification of gamma interferon, a lymphokine with potent

priming activity on macrophages. Many triggering stimuli

have also been identified, such as LPS, heat-killed











Listeria monocytogenes, double-stranded RNA, maleylated and

acetylated proteins, muramyl dipeptide, and lymphokines.

The interaction of many activating stimuli with

macrophages is poorly understood. Many laboratories have

reported binding parameters for gamma interferon on various

cell types, but only two groups have reported studies using

murine macrophages, with highly disparate results (4,27).

Studies on murine macrophages are critical due to the fact

that these cells have been extensively used to rigorously

characterize the phenomenon of activation for tumor cell

killing. The nature of the interaction of triggering

stimuli with macrophages is largely unknown. Specific

receptors have been characterized for maleylated proteins

(74), and extensive efforts to identify receptors for LPS

have been undertaken (117), but triggering of primed

macrophages for tumoricidal activation remains poorly

understood at the subcellular level.

The studies to be described here examine the

interactions of various agents involved in macrophage

activation with murine macrophages, either directly (in the

case of MuIFN-gamma interaction) or indirectly (in the case

of triggering stimuli), and finally describe the first

steps toward isolation and purification of the MuIFN-gamma

receptor for further molecular studies. These results

provide insights into the cell surface interactions

involved in the induction of macrophage activation, and








38

provide the groundwork for the purification of a specific

receptor for murine gamma interferon. The latter goal will

permit the production of antibodies to the receptor and its

molecular characterization, which represents the next

vertical step in the understanding of the induction of

macrophage activation for tumor cell killing.
















CHAPTER II

THE INTERACTION OF MURINE GAMMA INTERFERON
WITH MACROPHAGES


Introduction

Gamma interferon has multiple effects on macrophages

which, taken together, result in an overall enhancement of

their immune and host defense functions. These include the

up-regulation of Fc receptors and histocompatibility

antigens, the enhancement of microbicidal activity, and the

induction of nonspecific tumoricidal activity (reviewed in

161,192,203). In most instances, gamma interferon does not

fully activate macrophages for tumor cell killing, but

rather primes them to respond to a second, triggering

signal, most commonly provided by bacterial

lipopolysaccharide (LPS) (132). Macrophages can be

activated with gamma interferon alone under certain

conditions; this phenomenon is particularly dependent upon

the strain of mouse used for study (134). Therefore, the

physiological relevance of priming and triggering remains

controversial.

The first step toward understanding the mechanisms)

by which gamma interferon mediates its pleiotropic effects

is the characterization of a cell surface receptor for this

39











lymphokine. Specific, high affinity receptors for gamma

interferon have been demonstrated on numerous cell types of

both human (8,16,26,50,98,111,128,143,156,186,187) and

mouse (4,27,89,200) origin. The majority of studies on

IFN-gamma interactions with macrophages or macrophage-like

cell lines have been with human cells (26,50,128,143).

Receptors for IFN-gamma on mouse macrophages have been

described, although there is little agreement on the

binding parameters derived in the two extant studies

(4,27). The substantial inconsistencies in these data

necessitated the studies described here, under conditions

which the significant endocytic activity of macrophages

were known to be minimized. Murine gamma interferon was

labelled to high specific activity with 12sI and used to

characterize the interaction of IFN-gamma with various

mouse macrophage populations. The results indicated the

presence of a single class of specific, high affinity

binding sites for IFN-gamma on macrophages. Populations of

macrophages varying in their responsiveness to IFN-gamma

for activation did not appear to differ significantly in

receptor expression, suggesting that regulation of response

to activating signals occurs at a level distinct from

receptor interaction with the priming ligand.











Materials and Methods

Media and reagents. All experiments were performed in

HEPES-buffered, modified Eagle's medium (HMEM). HMEM

consisted of modified Eagle's minimum essential medium

(Auto-Pow MEM) prepared with 2 mg/ml sodium bicarbonate, 2

mM glutamine (all from Flow Laboratories, McLean, VA), 100

U/ml penicillin G potassium, 100 ug/ml streptomycin (both

from Pfizer, Inc., New York, NY), and 15 mM N-2-

hydroxyethylpiperazine-N'-2-ethane-sulfonic acid (HEPES;

Sigma Chemical Co., St. Louis, MO). For experiments, HMEM

was supplemented with 10% v/v fetal bovine serum (FBS;

Sterile Systems, Inc., Logan, UT). Heparin sodium (O'Neal,

Jones, and Feldman, St. Louis, MO) was used at 2 U/ml in

HMEM for bone marrow and peritoneal harvest. For bone

marrow culture, HMEM was supplemented with 15% v/v L cell-

conditioned medium, 10% v/v FBS, and 5% v/v horse serum

(Flow Laboratories, McLean, VA). L cell-conditioned medium

was obtained from cultures of NCTC L929 cells grown at

confluency for 4 days in HMEM/10% FBS, and filtered (0.45

Um) before use). All media tested negative for the

presence of endotoxin by the Limulus amebocyte lysate (LAL)

(Associates of Cape Cod, Inc., Woods Hole, MA) assay (97)

at a sensitivity of 0.07 ng/ml. Purified, lipid A-rich

fraction II of phenol-extracted bacterial

lipopolysaccharide (LPS) from Escherichia coli 0111:B4

(116) was a gift from Dr. D. C. Morrison, Department of











Microbiology, University of Kansas School of Medicine,

Kansas City, KS.

Bone marrow culture. Macrophages were obtained from

bone marrow cultured as described by Leung et al. (96) as

modified from Meerpohl et al. (107). Bone marrow was

harvested from the tibiae and femora of 6-9 week old

CaH/HeN mice (Charles River Laboratories, Kingston, NY) and

cultured in bacterial grade 100 mm2 Petri dishes (Lab-Tek

Products, Naperville, IL) at 370C, in an atmosphere

containing 5% C02. Bone marrow was seeded at 4 x 106

nucleated cells/dish in 15 ml of bone marrow culture

medium. On day 7, 5 ml of the same medium were added to

each dish. Cells were harvested for use between days 11 to

13 of culture using ice cold phosphate-buffered saline

(PBS). Differential analysis of cytocentrifuge prepara-

tions stained with Diff-Quik (Harleco, Gibbstown, NJ)

showed that populations of cells derived in this manner

consistently contained 97-99% macrophages.

Peritoneal macrophages. Peritoneal exudate cells were

harvested by lavage with HMEM containing 2 U/ml heparin.

For inflammatory exudates, mice were injected intraperi-

toneally with 1.5 ml 10% protease peptone broth three days

prior to harvest (130). Cells were incubated on tissue

culture-grade 96-well plates (200,000/well) for 2 hours at

370C for adherence, and nonadherent cells removed by

washing twice before initiating assays.











Interferon. Mouse gamma interferon (MuIFN-gamma) was

produced by E. coli into which an expression vector had

been introduced that contained the MuIFN-gamma coding

sequence (54,93). This reagent was a gift from Genentech,

Inc. (South San Francisco, CA) (lot #1551/43) and Schering-

Plough Pharmaceuticals (Bloomfield, NJ) (lots 18501-131

PF28-33 and 6-M6-01). The specific activity of the

Genentech interferon was 5 x 106 U/ml and of the Schering

interferons, 2 x 106 U/ml, as measured by the ability of

each preparation to protect L929 cells against plaque

formation by vesicular stomatitis virus (17).

Radioiodination of MuIFN-gamma. Gamma interferon was

radiolabelled with 1251 using chloramine T (71). Six pg of

interferon were diluted in 10 pl with 0.5 M sodium

phosphate buffer, pH 7.5, and combined with 3-5 pl (300-500

pCi) of carrier-free Na 1251 (15 mCi/pg; Amersham Corp.,

Arlington Heights, IL) and 10 pl 5 mg/ml chloramine T

(Sigma) for 60 seconds at 250C. The reaction was stopped

by the addition of 10 pi 10 mg/ml sodium metabisulfite and

10 pl 70 mg/ml potassium iodide. After adding 10 pl 20

mg/ml bovine serum albumin (BSA) and NaCI (4 M) to 1 M

final concentration, the mixture was desalted on a 5 ml

column of Bio-Gel P6-DG (Bio-Rad, Richmond, CA) that had

been equilibrated with 10 mM Tris-HCl, pH 7.5, 1 M NaCI,

containing 2 mg/ml BSA. The labelled preparations were

stored at 40C following addition of 2-mercaptoethanol to a











final concentration of 0.1% (v/v). Preparations were

tested for biological activity by measuring the ability to

prime macrophages for tumor cell killing (see below).

Assay for macrophage priming and activation. Priming

was defined as acquisition by macrophages of the capacity

to respond to a second, triggering signal (in this system,

LPS) by developing cytolytic activity for tumor cells. To

quantify priming activity (130), monolayers of bone marrow-

derived macrophages (5 x 104/well in 96-well microtiter

plates) were incubated at 37C with different concentra-

tions of MuIFN-gamma in the absence or presence of 1 ng/ml

LPS, for the times indicated in the text. After the

incubation period, monolayers were washed with HMEM/10%

FBS. s5Cr-labelled (Na251CrO4, 300 mCi/mg; Amersham Corp.,

Arlington Heights, IL) P815 mastocytoma cells were then

added in 200 Ul HMEM/10% FBS as targets (2 x 104/well).

After 16 hours at 37C, the upper 100 Ul of supernate were

removed and analyzed in an automatic gamma spectrometer to

determine the extent of 51Cr release. Cytotoxicity was

expressed as % specific 51Cr release as calculated using

the formula:

experimental cpm-spontaneous release cpm
total releasable cpm-spontaneous release cpm

Total releasable cpm was determined after freeze-thawing

under hypotonic (50% distilled water) conditions. Spon-

taneous release was that which occurred when target cells











were held throughout the incubation period on unstimulated

monolayers of macrophages.

Assay for binding of 125I-MuIFN-gamma. Macrophage

monolayers were prepared exactly as described above. After

allowing cells to adhere, they were washed and exposed to

50 pl of 100 mM NaN3 in HMEM/10% FBS for 15 minutes. In

pilot studies, macrophages that were exposed to 100 mM NaN3

for 15 minutes and then maintained in 50 mM NaN3 stopped

all detectable nonspecific endocytosis (using horseradish

peroxidase as a marker). The cells were not irreversibly

damaged, as full endocytic activity was recovered in NaN3-

free medium. An equal volume containing 1201-MuIFN-gamma

was next added to wells in the presence or absence of a

100-fold excess of unlabelled MuIFN-gamma. This

concentration of unlabelled MuIFN-gamma inhibited binding

by > 90%. Incubation at 37C for 2 hours was followed by

three rapid washes in ice-cold HMEM/10% FBS. Cells were

then solubilized in 2% NP-40, harvested, and lysates were

analyzed in an automatic gamma spectrometer. In some

experiments, cells were fixed with 0.1% paraformaldehyde in

phosphate-buffered saline to stop endocytic activity prior

to the binding assay. In these cases, NaN3 was omitted

from the assay.











Results

Radioiodination of murine gamma interferon. Purified,

recombinant-derived murine gamma interferon was radio-

labelled with '2I using chloramine T. Specific activities

ranged from 10-60 pCi per Ug protein. Greater than 50% of

the original biological activity was retained in these

preparations when tested for priming activity for

tumoricidal function. Alternative labelling techniques

resulted in lower specific activities with no significant

improvement in recovery of biological activity. Desalting

of unlabelled IFN-gamma indicated that significant losses

of activity occurred on the column. These losses were

overcome by performing the desalting procedure in the

presence of 1 M NaCl. Figure 2-1 represents an

autoradiograph of 125I-MuIFN-gamma analyzed by sodium

dodecylsulfate polyacrylamide gel electrophoresis (SDS-

PAGE). Two molecular species are evident: a major band

migrating with an apparent molecular mass (Mr) of 16,000

daltons, and a less predominant band of Mr 32,000 daltons.

The relative proportions of these species were unaffected

under reducing versus nonreducing conditions. For the

purposes of calculating binding parameters, it will be

assumed that the monomer species (Mr = 16,000) was bound.

Establishment of conditions to control macrophage

endocytic activity. Prior to attempting binding

experiments with 125I-MuIFN-gamma, it was necessary to










NR R


























M p I







Figure 2-1. SDS-PAGE of 125I-MuIFN-gamma.
125I-MuIFN-gamma (100,000 CPM/lane) was subjected
to electrophoresis on a 15% polyacrylamide resolving
gel under reducing (R) or non-reducing (NR)
conditions. Gels were stained with Coomassie blue for
localization of molecular weight markers, and dried
under heat and vacuum. Autoradiography using Kodak
XAR-5 film at -70oC revealed two species of
approximately 16,000 (monomer; M) and 32,000 dimerr;
D) daltons.











determine conditions under which endocytic function, and

therefore internalization of the ligand, could be

controlled. Using horseradish peroxidase (HRP) as a marker

of fluid-phase endocytosis, it was found that, at 370C,

high concentrations of sodium azide (NaN3) were required to

completely inhibit HRP uptake by bone marrow-derived

macrophages. Such NaN3 concentrations (up to 50 mM) were

not irreversibly toxic to the macrophages, as removal of

the NaN3 resulted in a rapid recovery of endocytic function

to control rates (data not shown).

Internalization of 25aI-MuIFN-gamma was directly

estimated by acid-stripping experiments. Bone marrow-

derived macrophages were exposed to 125I-MuIFN-gamma in the

presence or absence of a 100-fold excess of unlabelled IFN-

gamma to determine specific binding to these cells under

various conditions. At the end of 2 hours at 370C, the

cells were washed and treated with 0.2 M acetic acid, pH

2.5, containing 0.5 M NaCl for 5 minutes at 40C to remove

bound IFN-gamma remaining at the cell surface (59). The

cells were then lysed with 2% NP-40 to harvest remaining

(internalized or acid-resistant) counts. Table 2-1

indicates that under normal conditions, approximately two-

thirds of the bound IFN-gamma was internalized or resistant

to stripping. However, if the cells were treated with NaN3

before and during the binding assay, or if they were fixed


















Table 2-1. Inhibition of Internalization of 125I-MuIFN-
gamma by Macrophages



% Acid- % Acid-
Treatment Sensitive CPMd Resistant CPM


medium alone 33.9 66.1

PFb 80.2 19.8

NaN3C 77.9 22.1



aBone marrow culture-derived macrophages were allowed
to adhere to tissue culture wells for 2 hrs at 370C,
washed, and exposed to 100 ng/ml 125I-MuIFN-gamma 10
ug/ml unlabelled IFN-gamma for 2 hrs at 370C. All
incubations were in HMEM/10% FBS.

bImmediately prior to addition of 125I-MuIFN-gamma,
macrophages were fixed with 0.1% paraformaldehyde (PF) in
phosphate-buffer saline for 5 min.

cFifteen minutes before addition of 125I-MuIFN-gamma,
cells were exposed to sodium azide (NaN3) in HMEM/10% FBS
at a concentration of 50 mM; the presence of NaN3 was
maintained throughout the experiment.

dTo remove surface-bound (but not internalized) CPM,
macrophages were exposed to 0.2 M acetic acid, pH 2.5
containing 0.5 M NaCI for 5 min. at 40C, and washed once
with the same solution. Remaining (resistant) CPM were
recovered by solubilization of the cells with 2% NP-40.











with 0.1% paraformaldehyde, only 20% of the bound ligand

was acid-resistant. The complete lack of HRP internal-

ization with NaNa, coupled with the assumption that

paraformaldehyde-fixed cells are not endocytically active,

indicated that internalization of 125I-MuIFN-gamma was

inhibited to the greatest possible extent under these

conditions.

Binding parameters of 125I-MuIFN-gamma to macrophages.

Under conditions designed to minimize endocytosis of the

ligand, macrophages were incubated with 125I-MuIFN-gamma

(Genentech) to determine binding parameters for this

ligand. Time course experiments at 370C indicated that

binding equilibrium was gradually achieved by 2 hours of

incubation. Nonspecific binding, i.e., that which bound in

the presence of a 100-fold excess of unlabelled MuIFN-

gamma, remained unchanged over the same time, and was less

than 10% of the total MuIFN-gamma bound by the cells

(Figure 2-2). All subsequent binding experiments were

performed for 2 hours at 370C. The specificity of 125I-

IFN-gamma binding was further ascertained by incubating the

radioligand with macrophages in the presence of increasing

concentrations of unlabelled MuIFN-a, -0, or -gamma. Only

the MuIFN-gamma competed effectively for binding of 125I-

MuIFN-gamma, while the Type I IFN's (a and 3) were

ineffective, even at concentrations up to 2500 U/ml (Figure

2-3).











o Total
o Nonspecific
* Specific


DURATION OF EXPOSURE (minutes)


Figure 2.2.


Time course of binding of 1251-MuIFN-gamma to
macrophages.


Bone marrow culture-derived macrophages were allowed
to adhere to tissue culture wells for 2 hr, exposed to
NaN3 in HMEM/10% FBS as described in Materials and
Methods, and then exposed to 10 U/ml 125I-MuIFN-gamma
for the times indicated on the abscissa at 370C. At
those times, monolayers were washed rapidly with three
ice-cold volumes of HMEM/10% FBS and solubilized with
2% (w/v) NP-40. Nonspecific binding (that which bound
in the presence of a 100-fold excess of unlabelled
MuIFN-gamma) was subtracted from total binding (that
which bound in the absence of unlabelled MuIFN-gamma)
to calculate specific binding.


1000



800














z 80-



o 60-
600
0
LL 40-
40 a-IFN
o o) -IFN
Sa y -IFN
M 20-


I I I I
10 100 1000 10000

UNLABELLED IFN (U/ml)

















Figure 2-3. Specificity of binding of 125I-MuIFN-gamma to
macrophages.

Bone marrow culture-derived macrophages were adhered
and treated with NaN3 as in Figure 2-2. Cells were
then exposed to a constant concentration of 125s-
MuIFN-gamma (50 U/ml) in the presence of the indicated
concentrations of a-, P-, or gamma-interferons for 2
hr at 370C. Binding was determined after washing and
solubilization with 2% NP-40.











Binding parameters were determined on bone marrow-

derived macrophages under the conditions outlined above.

Increasing concentrations of 125I-MuIFN-gamma (Genentech)

were incubated with macrophage monolayers in the absence or

presence of a 100-fold excess of unlabelled IFN-gamma.

Figure 2-4 represents an equilibrium saturation binding

isotherm for macrophages. Specific binding was saturable

at ligand concentrations above 1 nM (75 U/ml for this

preparation). Scatchard analysis of the data indicated

that there was a single class of approximately 11,500

binding sites per cell, with a dissociation constant (Kd)

of 3.8 x 10-10 M (28 U/ml for this preparation).

Binding of 125I-MuIFN-gamma to modulated macrophage

populations. The ability to quantify receptor expression

on macrophages that differ in their responsiveness to

IFN-gamma provided the opportunity to determine whether

these differences were mediated at the level of the

receptor. Two models of macrophage responsiveness were

examined. In the first, bone marrow-derived macrophages

were pre-treated with a Type I interferon (in these

experiments, MuIFN-0), which has been shown to signifi-

cantly reduce the capacity of these macrophages to respond

to IFN-gamma and bacterial lipopolysaccharide (LPS) for

activation for tumor cell killing (129). Macrophages that

had been exposed to IFN-P under conditions in which their

response to IFN-gamma and LPS was substantially diminished














I (0
H 0>
0 -,-l

0



0 0
tI.4U
) 4 o)

a 0 0






0 0
,4 o 4
1 o o *


o o 9 -rI
*/ o> (


* c H 4- -r
cM CM 0 (
a 4-
I Z l 1 I 01

Wu a 4-4 -.rI
U-
COi x ( /id/Iow) 1/98 N. to
I 0 .*rl
o0-4 I C14

I 0 a
z 0) ,-) 44

d- R
\0 a i







\a .0

: r.4 (*r
S *rl C



,- r*-I a
k- 04
*) -00









0 VQ I
\I *H
Of i E B*rl











r (
-0 0








HH











were examined for their ability to bind 123I-MuIFN-gamma.

MuIFN-0-treated macrophages bound slightly less 125I-MuIFN-

gamma (92%) than control cells, but the difference was

minimal (Figure 2-5). No significant difference in

affinity of binding was observed.

The second model of variation in macrophage

responsiveness to IFN-gamma and LPS derives from studies of

peritoneal exudate-derived macrophages. Macrophages

derived from the peritoneal cavities of mice that have

previously been injected with a sterile irritant (in this

case, 1 ml fetal bovine serum containing 0.1 mg/ml LPS) are

consistently more responsive to activating agents (IFN-

gamma and LPS) than peritoneal macrophages from

unmanipulated mice (Table 2-2). However, when these

macrophages were compared for their IFN-gamma binding

characteristics, no differences were observed either in the

number of binding sites or in the affinity of binding

(Table 2-2).

Binding of 125I-MuIFN-gamma to primed and activated

macrophages. To determine if the state of activation of

macrophages affects IFN-gamma binding, bone marrow-derived

macrophages were treated with IFN-gamma (primed) or IFN-

gamma and LPS (activated) and tested for IFN-gamma binding

(Figure 2-6). Prior to performing the binding assay, the

cells were incubated with three successive 1.5-hour washes

to allow any bound IFN-gamma to be internalized or
































B(fmol)
2 4 6
125I-MuIFN- y OFFERED(nM)


Figure 2-5.


Effect of IFN-O pretreatment on specific
binding of 125I-MuIFN-gamma.


Bone marrow culture-derived macrophages were treated
for 24 hr at 370C with 100 U/ml IFN-O (o) or medium
(*). Monolayers were washed and specific binding of
125I-MuIFN-gamma (Schering) determined as in Figure 2-
4. Binding parameters were determined by Scatchard
analysis (inset). Control macrophages bound 7.55
femtomoles 125I-MuIFN-gamma with a Kd of 4.15 nM,
while IFN-0-treated cells bound 6.96 femtomoles with a
Kd of 5.14 nM.




















Table 2-2. Interferon Receptor Expression on Resident vs.
Inflammatory Peritoneal Macrophages



ED5o (U/ml)a

IFN IFN Receptors/
alone + LPS Cellb Kd


Resident 42.0 1.02 35669 3044 4.36 x 10-9M

Inflammatory 6.1 0.16 34916 2189 3.15 x 10-9M



aEDso refers to the concentration of gamma interferon,
in the absence or presence of 3 ng/ml LPS, that induces 50%
of the maximal tumoricidal response. Peritoneal exudate
cells were harvested by lavage from C57BL/6 mice that were
injected 24 hours earlier with 1 ml fetal bovine serum
containing 0.1 ng/ml LPS (inflammatory) or were
unmanipulated (resident). Cells were allowed to adhere to
96-well tissue culture plates (200,000 macrophages/well)
and washed to remove nonadherent cells. stimuli were added
for 6 hours, after which the monolayers were washed and
exposed to s5Cr-labelled P815 target cells for 16 hours.
bSpecific binding of 125I-MuIFN-gamma (Schering) was
determined on monolayers in parallel with those used for
cytotoxicity determinations.













2 -








ZI








0 2 4 6 8
5 10
125
125 MuFN- OFFERED (nM)


















Figure 2-6. Binding of 1251-MuIFN-gamma to primed and
activated macrophages.

Bone marrow culture-derived macrophages were treated
for 24 hr with either medium alone (M), 20 U/ml MuIFN-
gamma (o), or 20 U/ml MuIFN-gamma + 1 ng/ml LPS (e).
After three successive 1.5-hr incubations (370C) in
HMEM/10% FBS to remove and allow internalization of
the MuIFN-gamma, macrophages were fixed with 0.1%
paraformaldehyde and incubated with 25"I-MuIFN-gamma
(Schering) to determine specific binding parameters.
Femtomoles bound and dissociation constants were
calculated: control, 7.80, 4.75 nM; primed, 4.17,
3.53 nM; activated, 4.01, 5.73 nM.











dissociated. Under these conditions, both primed and

activated macrophages bound 50% less IFN-gamma than

untreated controls. The reduction in binding was in

receptor number, and no change in the affinity of binding

was observed.


Discussion

The studies described here confirm the existence of

specific, high affinity binding sites for mouse gamma-

interferon on macrophages, and document the expression of

these sites on different macrophage populations. These

binding sites are believed to be receptors for IFN-gamma

that represent the first step in the ultimate functional

modulations of macrophages that are mediated by this

ligand. While the complex nature of the pleiotropic

responses of macrophages to IFN-gamma is not yet

understood, the characterization of a specific receptor for

this ligand is the necessary first step toward delineating

the signals required for these varied responses.

Purified mouse gamma interferon, derived from

Escherichia coli into which an expression vector containing

the cDNA (54) had been introduced, was labelled with 1251

for use as a radioligand in binding experiments. Use of

this radioligand as a priming stimulus in the induction of

macrophage tumoricidal activity indicated that greater than

50% of the biological activity was retained. The specific











activity of the labelled IFN-gamma (=10 pCi/pg) was quite

high and similar to that reported by others (4,27,200).

The only two extant studies of IFN-gamma interaction

with mouse macrophages have reported contrasting results

(4,27). The data presented here support the work of

Celada, et al. (27), i.e., that mouse macrophages, derived

from bone marrow culture, possess a single class of

relatively high affinity (0.4-4 nM, depending upon the

MuIFN-gamma used) binding sites for IFN-gamma. Celada, et

al. (27) reported a single class of IFN-gamma binding sites

with a Kd of 1 x 10-8 M. In a later, more extensive

report, however, Aiyer et al. proposed that macrophages

possessed two classes of IFN-gamma receptors, one of

intermediate affinity (3 x 10-9 M), and one of high

affinity (9 x 10-11 M) (4). The latter was claimed to be

the functional receptor, as its affinity was more

consistent with physiologically active concentrations of

the ligand. The results of the latter report were derived

under conditions where internalization of the ligand was

likely to have occurred; this is a more likely explanation

of the curvilinear Scatchard plot derived in these studies,

particularly in light of the extraordinary endocytic

activity of macrophages (179). A recent report by Celada

and Schreiber (29) documents the internalization of IFN-

gamma by mouse macrophages; the data in Table 2-1 supports

this finding.











The binding parameters derived in this study varied

according to the preparation of MuIFN-gamma used. The

Genentech preparation consistently yielded higher affinity

estimates, with lower nonspecific binding. The Kd derived

from the Schering preparations was roughly an order of

magnitude lower (mid-nanomolar range) than the Genentech

(Figures 2-5, 2-6, and Table 2-2). Differences of this

magnitude are not unusual among IFN-gamma preparations

(93). The specific antiviral activities of these

preparations were reported by the suppliers as 2 x 107 and

1 x 10' U/mg for Genentech and Schering, respectively. In

our hands, these activities were 5 x 106 and 2 x 106 U/mg,

respectively. Differences in biological assay systems may

have been responsible for these disparities.

For functional comparison, the 125I-MuIFN-gamma used

to derive the binding parameters in Figure 2-4 was directly

assayed for biological activity (macrophage priming). The

results indicate that priming can occur at concentrations

well below half-maximal occupation of receptors (Kd = 28

U/ml). However, the concentrations of MuIFN-gamma required

for priming are interdependent with the amount of

triggering agent supplied as well (130). In addition, in

strains of mice where MuIFN-gamma activates alone (or in

the presence of undetectable amounts of a triggering

agent), the concentrations required are low in comparison

to the receptor affinity, suggesting that only a small











number of receptors need to be occupied for a significant

functional response. The hypothesis has been advanced that

multiple cycles of internalization of receptor-ligand

complexes are required for the induction of the tumoricidal

response (29). This would explain the fact that few

receptors need to be occupied, as measured under static

conditions, to elicit a complex response.

Studies of IFN-gamma receptor expression on different

macrophage populations indicated that modulation of

receptors is not likely the basis for differential

responsiveness to IFN-gamma for tumoricidal activation.

While some down-regulation of receptors on macrophages was

apparent after pretreatment with Type I interferons, this

change seemed insufficient to explain the drastic loss of

responsiveness of the treated cells. The increased

responsiveness of macrophages from inflammatory sites also

appeared to be mediated at a level other than the IFN-gamma

receptor. Resident and inflammatory macrophages were

comparable on the basis of receptor number and affinity for

IFN-gamma. These two models indicate that regulation of

macrophage responses to IFN-gamma and triggering stimuli

for tumoricidal activation is either mediated at a post-

receptor level or at the level of the triggering stimulus

(in this case, LPS). To date, there has been no evidence

for regulation of IFN-gamma receptor expression by any

stimuli that are known to affect macrophage functions.











The homologous ligand, IFN-gamma, does appear to

affect the expression of its receptor. Macrophages that

were either primed (treated with IFN-gamma alone) or

activated (treated with IFN-gamma and LPS) both exhibited

decreased binding of MuIFN-gamma. Binding of the ligand to

these populations was reduced by about 50%. This "down-

regulation" was specifically induced by IFN-gamma, and was

independent of the state of activation. The decreased

binding is not likely to be due to occupation of receptors

by the MuIFN-gamma used to stimulate the cells, since 1) a

very low concentration of MuIFN-gamma (20 U/ml) was used

for priming and 2) the cells were washed with three

successive 1.5-hour washes prior to the binding assay.

These results are at odds with a report by Celada and

Schreiber (29) that IFN-gamma-treated macrophages are not

affected in their receptor expression. However, no data

was shown in this report; direct comparison with these data

were therefore impossible. The 4.5-hour incubation here,

after removal of the ligand, probably allows

internalization of occupied receptors that are not recycled

in the absence of the ligand. Similar homologous ligand-

induced receptor down-regulation has been reported for

human IFN-a (22), platelet-derived growth factor (65), and

epidermal growth factor (3), and is relatively common

phenomenon (142).











These studies corroborate the presence of a specific,

high affinity, single class of receptors for gamma

interferon on mouse macrophages. They refute the previous

claim of two distinct binding sites for MuIFN-gamma on

macrophages (4), and indicate binding sites with an

affinity at least 10-fold higher than that reported by

Celada et al. (27). The examination of macrophage

populations with altered responsiveness to activating

stimuli indicated that IFN-gamma receptor expression is

most likely not a major point of regulation for differences

in capacity for activation. Finally, MuIFN-gamma appears

to down-regulate its own receptor, although the functional

consequences of such down-regulation are not readily

apparent.


















CHAPTER III

INTERACTIONS OF PRIMING AND TRIGGERING STIMULI
WITH MACROPHAGES AS REVEALED BY CELL SURFACE PROTEOLYSIS


Introduction

Macrophages can be activated to kill neoplastic cells

as the result of a series of steps induced by a sequence of

stimuli (68,108,150). Quiescent macrophages first reach an

inflammatory stage (1,68,108), then become primed to

respond to a final triggering signal to become fully

cytolytic. While stimuli inducing the inflammatory stage

are largely undefined, priming and triggering agents have

been well-documented. The only known priming agents are

the interferons. Of these, gamma interferon (IFN-gamma) is

much more potent for priming than a/P interferons when they

are compared on the basis of antiviral units (131).

Triggering stimuli can be provided by a number of

substances, including bacterial lipopolysaccharide (LPS)

(5,130), heat-killed Listeria monocytogenes (HKLM) (165),

and polyinosinic-polycytidylic acid (poly I:C) (183,185).

When used at high concentrations, these agents can activate

macrophages fully; however, direct activation of this

nature still occurs by a two-step process, in that the







66

triggering agents induce a/0 interferon production by the

macrophage, which then primes in an autocrine manner (166).

While the phenomenon of macrophage activation is well-

characterized (1), little is known about its induction at

the subcellular level.

The first step in the priming of macrophages by murine

IFN-gamma (MuIFN-gamma) most likely requires the engagement

of a specific receptor for this lymphokine (27,151,163).

Little is known, however, about how the various triggering

agents provide their signal for the induction of cytolytic

activity. The results presented here will demonstrate that

binding of 125I-MuIFN-gamma and the induction of activation

are affected differently by prior treatment of macrophages

with various proteolytic enzymes. The data indicate that,

in addition to engagement of the MuIFN-gamma receptor,

other cell surface structures are required for activation.

Proteolysis also differentiated between two means of

triggering primed macrophages, and dissociated two induced

proteins, previously thought to be specific for the

triggering response, from the triggering phenomenon.



Materials and Methods

Media and reagents. All experiments were performed in

HEPES-buffered, modified Eagle's medium (HMEM). HMEM

consisted of modified Eagle's minimum essential medium

(Auto-Pow MEM) prepared with 2 mg/ml sodium bicarbonate, 2











mM glutamine (all from Flow Laboratories, McLean, VA), 100

U/ml injectable penicillin G potassium, 100 pg/ml

injectable streptomycin (both from Pfizer, Inc., New York,

NY), and 15 mM N-2-hydroxyethylpiperazine-N'-2-ethane-

sulfonic acid (HEPES; Sigma Chemical Co., St. Louis, MO).

For experiments, HMEM was supplemented with 10% v/v fetal

bovine serum (FBS; Sterile Systems, Inc., Logan, UT).

Heparin sodium (O'Neal, Jones, and Feldman, St. Louis, MO)

was used at 2 U/ml in HMEM for bone marrow harvest. For

bone marrow culture, HMEM was supplemented with 15% v/v

L cell-conditioned medium, 10% v/v FBS, and 5% v/v horse

serum (Flow Laboratories, McLean, VA). L cell conditioned

medium was obtained from cultures of NCTC L929 cells grown

at confluency for 4 days in HMEM/10% FBS, and filtered

(0.45 pm) before use. All media tested negative for the

presence of endotoxin by the Limulus amebocyte lysate (LAL)

(Associates of Cape Cod, Inc., Woods Hole, MA) assay (97)

at a sensitivity of 0.07 ng/ml. Purified, lipid A-rich

fraction II of phenol-extracted bacterial lipopolysac-

charide (LPS) from Escherichia coli 0111:B4 (116) was a

gift from Dr. D. C. Morrison, Department of Microbiology,

University of Kansas School of Medicine, Kansas City, KS.

Heat-killed Listeria monocytogenes was a gift from Dr. R.

D. Schreiber, Washington University, St. Louis, MO.

Polyinosinic-polycytidylic acid was from Calbiochem (San

Diego, CA).











Bone marrow culture. Macrophages were obtained from

bone marrow cultured as described by Leung et al. (96) as

modified from Meerpohl et al. (107). Bone marrow was

harvested from the tibiae and femora of 6-9 week old

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

cultured in bacterial grade 100 mm2 Petri dishes (Lab-Tek

Products, Naperville, IL) at 370C, in an atmosphere

containing 5% CO2. Bone marrow was seeded at 4 x 106

nucleated cells/dish in 15 ml of bone marrow culture

medium. On day 7, 5 ml of the same medium were added to

each dish. Cells were harvested for use between days 11 to

13 of culture using ice cold phosphate-buffered saline

(PBS). Differential analysis of cytocentrifuge prepara-

tions stained with Diff-Quik (Harleco, Gibbstown, NJ)

showed that populations of cells derived in this manner

consistently contained 97-99% macrophages.

Interferon. Recombinant mouse gamma interferon

(MuIFN-gamma) was produced by E. coli into which an

expression vector had been introduced that contained the

MuIFN-gamma coding sequence (54). This reagent was a gift

either from Genentech, Inc. (lot #1551/43; South San

Francisco, CA) or from Schering-Plough Corp. (lots #18501-

131 PF 28-33 and 6-M6-01; Bloomfield, NJ).

Proteolytic enzymes. Pronase (protease Type XIV from

Streptomyces griseus), trypsin (Type III, from bovine

pancreas), and chymotrypsin (Type VII, TLCK-treated, from











bovine pancreas) were all obtained from Sigma. The

specific caseinolytic activities determined for these

enzymes were 3.4 U/mg, 5.7 U/mg, and 6.4 U/mg, respec-

tively, where a unit was defined as the ability to produce

1 Umole of tyrosine equivalents from the substrate (0.65%

casein, w/v) per minute at 37C.

Radioiodination of MuIFN-qamma. MuIFN-gamma was

radioiodinated using either immobilized glucose oxidase-

lactoperoxidase (Enzymobeads; BioRad, Richmond, CA),

according to the manufacturer's instructions, or the

chloramine T method (71). The labeled preparations were

stored at 40C after addition of 2-mercaptoethanol to 0.1%

(v/v). Each radioiodinated preparation was tested for

biological activity by measuring its ability to prime

macrophages for tumor cell killing (see below). The

results were used to calculate the specific activity of

each preparation using unlabelled MuIFN-gamma as a

reference. Specific activities ranged from 10-50 uCi/pg

with greater than 50% recovery of biological activity.

Assay for macrophage priming and activation. Priming

was defined as acquisition by macrophages of the capacity

to respond to a second, triggering signal (in this system,

either LPS, heat killed Listeria monocytogenes (HKLM], or

poly I:C) by developing cytolytic activity for tumor cells.

To quantify priming activity (130), monolayers of bone

marrow culture-derived macrophages (5 x 104/well in 96-well











microtiter plates) were incubated at 370C with different

concentrations of MuIFN-gamma in the absence or presence of

the triggering stimulus, for the times indicated in the

text. After the incubation period, monolayers were washed

with HMEM/10% FBS. 51Cr-labelled (Na25sCrO4, 300 mCi/mg;

Amersham Corp., Arlington Heights, IL) P815 mastocytoma

cells were then added in 200 pl HMEM/10% FBS as targets (2

x 104/well). After 16 hr at 370C, the upper 100 pl of

supernate were removed and analyzed in an automatic gamma

spectrometer to determine the extent of 51Cr release.

Cytotoxicity was expressed as % specific 51Cr release as

calculated using the formula:


experimental cpm-spontaneous release cpm
x 100
total releasable cpm-spontaneous release cpm


Total releasable cpm were determined after freeze-thawing

under hypotonic (half distilled water) conditions.

Spontaneous release was that which occurred when target

cells were held throughout the incubation period on

unstimulated monolayers of macrophages.

Assay for binding of 125I-MuIFN-gamma. Macrophage

monolayers were prepared exactly as described above. After

allowing cells to adhere, they were washed and exposed to

50 pl of 100 mM NaN3 in HMEM/10% FBS for 15 minutes. In

pilot studies, macrophages that were exposed to 100 mM NaN3

for 15 minutes and then maintained in 50 mM NaNa stopped











all detectable nonspecific endocytosis (using horseradish

peroxidase as a marker) and recovered endocytic activity in

NaNa-free medium. An equal volume containing 125I-MuIFN-

gamma was next added to wells in the presence or absence of

a 100-fold excess of unlabelled MuIFN-gamma. In pilot

experiments, this concentration of unlabelled MuIFN-gamma

inhibited binding by >90%. Incubation at 370C for 2 hours

was followed by 3 rapid washes in ice-cold HMEM/10% FBS.

Cells were then solubilized in 1% SDS, harvested, and

lysates were analyzed in an automatic gamma spectrometer.

Detection of proteins induced by activating agents.

The induction of specific newly synthesized proteins in

macrophages treated with various stimuli was performed as

described by MacKay and Russell (103). Macrophage

monolayers were simultaneously exposed to stimuli (IFN-

gamma and triggering agents) and 20 uCi L-[35S] methionine

(Specific activity = 1060 Ci/mmol; Amersham, Arlington

Heights, IL) for 6 hr in methionine/leucine/lysine-free

HMEM + 10% fetal bovine serum. After labelling, monolayers

were washed three times in phosphate-buffered saline and

solubilized in a lysis solution composed of 9.5 M urea, 2%

(w/v) NP-40, 100 mM dithiothreitol, 2% w/v ampholytes, pH

3.5-10 (Ampholines, LKB Instruments, Gaithersburg, MD) and

protease inhibitors phenylmethylsulfonylfluoride and

pepstatin A. Lysates were immediately frozen and stored at

-70oC. Trichloroacetic acid-precipitable radioactivity was











determined, and equal counts (usually 1.5 x 106 CPM) were

analyzed by two-dimensional gel electrophoresis as modified

from O'Farrell (126). Samples were subjected to isoelec-

tric focusing in the first dimension, in 2.5 x 125 mm tube

gels containing 4.24% w/v acrylamide, 0.74% DATD, 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). After focusing for 7000 volt-hours,

first-dimension gels were equilibrated in SDS sample buffer

and frozen (-70oC) until used for the second dimension,

SDS-polyacrylamide gel electrophoresis. Tube gels were

subjected to SDS-PAGE on 160 mm wide x 280 mm long x 1.6 mm

thick resolving slab gels (10% w/v acrylamide with 0.27%

w/v bis-acrylamide). Electrophoresis was performed at 11

mA constant current, then processed for fluorography after

fixation in TCA. Processing was performed by immersing

gels in two successive baths of dimethylsulfoxide to

dehydrate them, followed by impregnation with 2,5-diphen-

yloxazole for 1.5 hours. After rehydration of gels to

remove DMSO, gels were dried and exposed to XAR-5

radiographic film (Kodak, Rochester, NY) for 1.5 x 106

CPM-days.



Results

Time course of macrophage activation for tumor cell

killing. Bone marrow-derived macrophages were exposed for

various times to various concentrations of MuIFN-gamma with











or without a constant amount of LPS (1 ng/ml) (Figure 3-1).

As expected, with either MuIFN-gamma or LPS used alone at

such a low concentration, no killing was observed. A

response of the macrophages to MuIFN-gamma in the presence

of 1 ng/ml LPS was observed. It was both dose and time-

dependent. At concentrations of 5 U/ml and above, stimu-

lation was maximal and the responses, therefore, indistin-

guishable. In each instance they began to develop after 2

hr and peaked by 6 hr. In contrast, at the lower concen-

tration of 1 U/ml, the response developed at the same rate,

but only after an initial lag period of 2 additional hr.

As expected (108), the tumoricidal capacity of all mono-

layers gradually decreased with continuing exposure to the

activating agents. All subsequent experiments were

performed with a 6-hr stimulus exposure to achieve the

maximal response at appropriate concentrations of stimuli.

Effect of proteases on macrophage activation for tumor

cell killing. Three proteolytic enzymes with different

specificities (trypsin cleaves after arginines and lysines,

chymotrypsin after aromatic amino acids, and Pronase is

broadly active, cleaving bonds between many different amino

acids) were investigated to determine whether or not they

would affect responsiveness of macrophages to MuIFN-gamma

and LPS. Macrophage monolayers were pre-treated with

either Pronase, trypsin, or chymotrypsin in serum-free

medium for the periods of time indicated on the abscissa.































4 8 12 16 20

DURATION OF EXPOSURE (hours)


Figure 3-1.


Time course of macrophage activation by MuIFN-
gamma in the presence of LPS.


Bone marrow culture-derived macrophages were seeded in
96-well tissue culture plates and allowed to adhere
for 2 hr at 370C. Monolayers were washed and MuIFN-
gamma was added at 1 (o), 5 (A), 10 (0), or 50 (*)
U/ml in the presence or absence of 1 ng/ml LPS for the
times indicated on the abscissa. Stimuli were
removed, the monolayers were washed, and 53Cr-labelled
P815 target cells were added for an additional 16 hr
to determine cytolytic activity by 51Cr release. No
killing was observed in the absence of LPS or with LPS
alone.











The monolayers were then washed and incubated for 5 min in

20% FBS to minimize residual proteolytic activity. i25j-

MuIFN-gamma was then added, with or without 1 ng/ml LPS.

Exposure to each of the proteases resulted in a reduction

in the capacity of macrophages to become activated to a

tumoricidal state (Figure 3-2). However, the enzymes had

different effects. All three had a marked, initial effect,

although Pronase caused the most pronounced reduction in

killing. After 30 min the magnitude of the cytolytic

response to activators continued to be reduced, but at a

different rate than that which had been observed earlier.

Effect of proteases on specific binding of 125I-

MuIFN-gamma. To determine if the reduction of macrophage

responsiveness to activating stimuli was due to decreased

binding of MuIFN-gamma, binding experiments were performed

in parallel with the functional studies. The same pools of

reagents and cells were used to optimize comparison between

the two kinds of experiments. Figure 3-3 shows the effects

of the three proteases on the capacity of macrophages to

bind 125-I-MuIFN-gamma. While trypsin and chymotrypsin had

had a significant effect on responsiveness (Figure 3-2),

these enzymes had no detectable effect on the binding of
12sI-MuIFN-gamma. Pronase-treated macrophages, on the

other hand, bound approximately 40% less interferon, and

most of the loss occurred within the first 30 min of

incubation. Equilibrium saturation binding analysis over a














x 80-
0O
0

S60-
.J
0

o 40-

0
-20-
w

QII I
30 60 90 120
DURATION OF PROTEOLYSIS (minutes)









Figure 3-2. Effect of proteases on responsiveness of
macrophages to MuIFN-gamma and LPS.

Bone marrow culture-derived macrophage monolayers were
prepared in serum-free medium and exposed to 50 pg/ml
Pronase (A), trypsin (o), or chymotrypsin (V) at 37C
for the indicated times. Based on the caseinolytic
activities of these enzymes, this represented 0.170,
0.285, and 0.320 U/ml, respectively. After protease
treatment, cells were washed and incubated in medium
containing 20% fetal bovine serum (FBS) for 5 min at
370C to minimize residual proteolytic activity. The
cells were then exposed to 50 U/ml MuIFN-gamma + 1
ng/ml LPS for 6 hr, washed to remove the stimuli, and
incubated for an additional 16 hr with 35Cr-labelled
P815 target cells to determine cytolytic activity.
Results are means 1 SEM from two separate
experiments (n = 4).













100



80



60



40


30 60 90
DURATION OF PROTEOLYSIS (minutes)


Figure 3-3.


Effect of proteases on specific binding of
25 I-MuIFN-gamma.


Macrophage monolayers were treated in parallel with
those from Figure 3-2 with 50 ug/ml Pronase (A),
trypsin (o), or chymotrypsin (V). Specific binding of
125I-MuIFN-gamma was then determined as described in
Materials and Methods, at a concentration of 50 U/ml
125MuIFN-gamma 5000 U/ml unlabelled MuIFN-gamma.


120











range of 125I-MuIFN-gamma concentrations indicated no

decrease in either the affinity of binding or the number of

binding sites as a result of trypsin treatment, even when

the concentration of trypsin was increased 10-fold (to 500

ug/ml) (data not shown).

Regeneration of responsiveness to activating stimuli

by Pronase-treated macrophages. To ensure that the effects

on responsiveness were not due to a trivial explanation,

e.g., a cytotoxic effect on the macrophages, experiments

were undertaken to determine if protease-treated monolayers

could regenerate their capacity to respond to activating

stimuli. Pronase was chosen for these experiments because

it had previously had the most extensive effect. As can be

seen in Figure 3-4, control macrophages began to express

cytolytic activity in response to MuIFN-gamma and LPS

stimulation after 4 hr, while monolayers pretreated with

Pronase became responsive to MuIFN-gamma and LPS after an

additional lag period of about 8 hr. At that time, treated

macrophages began to develop the capacity to kill at the

same rate and ultimately reached the same maximal level of

cytolytic activity that the untreated controls had.

Regeneration of binding of 12 I-MuIFN-gamma by prote-

ase-treated cells. Experiments were also performed to

determine if the binding capacity of Pronase-treated

macrophages could be regenerated. Macrophage monolayers

were treated as before, but with 500 ug/ml Pronase, an




















50-



U-


W 30-
a-

z
w

a-
0 -



12 16 20 24
TIME FOLLOWING PROTEOLYSIS (hours)











Figure 3-4. Regeneration of responsiveness of Pronase-
treated macrophages to MuIFN-gamma and LPS.

Macrophage monolayers were treated with 500 pg/ml
Pronase (w) or medium alone (o) for 2 hr, after which
they were washed and incubated in 20% FBS for 5 min.
Macrophages were immediately exposed to 50 U/ml
MulFN-gamma 1 ng/ml LPS in HMEM/10 FBS for the
times indicated on the abscissa. At those times,
monolayers were washed to remove stimuli, and 51Cr-
labelled P815 target cells were added to determine
cytolytic activity.











amount that was shown in pilot experiments to nearly

eliminate binding of MuIFN-gamma (10% of specific binding

on control cells; data not shown). After exposure to the

protease, cells were incubated for varying times in either

medium alone or medium that contained 20 ug/ml cyclohex-

imide, and specific binding was quantified at each time

point. As is shown in Figure 3-5, specific binding

capacity for MuIFN-gamma began to return after

approximately 2 hr and peaked by 4 hr. Recovery did not

occur in the presence of cycloheximide although macrophages

remained viable (trypan blue dye exclusion as the

criterion).

Status of trypsin-treated macrophages. Further

experiments were performed to determine the nature of the

trypsin-induced decrease in responsiveness of macrophages.

Macrophage monolayers were pretreated with 500 ug/ml (2.85

caseinolytic U/ml) trypsin for 1 hour, then exposed to a

combination of MuIFN-gamma and three different triggering

agents (LPS, HKLM, and poly I:C) to determine if the loss

of responsiveness was a generalized phenomenon or due to a

selective loss of responsiveness to triggering stimuli.

Dose-response cytotoxicity data are shown in Figure 3-6 for

each of the triggering agents in the presence of a constant

amount of MuIFN-gamma (10 U/ml). In each case, macrophages

responded to the triggering agents in a dose-dependent

manner. No killing was observed with the triggers alone




































2 4 6 8
TIME FOLLOWING PROTEOLYSIS (hours)


Figure 3-5.


Regeneration of binding of 125I-MuIFN-gamma by
Pronase-treated macrophages.


Macrophage monolayers were treated with 500 pg/ml
Pronase in serum-free medium for 2 hr at 370C. After
washing and incubation in 20% FBS for 5 min, cells
were allowed to recover in the absence (V) or presence
(o) of 20 ug/ml cycloheximide in HMEM/10% FBS for the
times indicated on the abscissa. Monolayers were then
washed and specific binding of 12sI-MuIFN-gamma
determined.
















































































0 0 0

JOI g 01tdIO3dS %


0
0



0
C')



o 1
- 0






o i


0


E


I


* -


C')
S
0
~
OS
-J
In Z
0
K;
0

IL)
0


*r.

(0

(0
Ol

I4

H



0
.74
to






0
rz

>1
A




4)
*P
0
*rl
ce

4)
0

0
.4
4 0


0>1











'-4
0 *








40
0
4J 4
0 )


















144
(144
H

o >





>10
P o

a .
0 1
1 *)
E 2







4P)14
IM
em a


Q.
a &





0)
'4.4 01


!>0
S0)
U
0C )
014
aP ^
7
0)4
44 )
44 X
MA P


*
a)
I


0)4


o I41P
-0 4 *
$ 0 0 0
*-0 P0 4 M

'- Eu in O : $
) 0i r-I 0
r. 0 H4) H 0Q 0 0
04- 0P k -*rl
a0 w 0 O r-1 m1
54 O3 .r- -r


r4.4
C Vrl 41 H A 0 tB

II(l .z


Oro 0 >r

d -0 *- 0
0,U >1 4
iC 01 r-4 4-) a)









o 0 I) 0
m I M W r*









g0 .,0 00
4 a) (d0M r-1 >v 4.)






4 rM 144
0 0,-H., 0











04)74$0 0) 0


0o 4000
4, o H o >






-3 0 *-









iOWC *r.U

0 i 0 ,4)10)00
HO*H ( 0
0 4. 0 $4 V
4) Id 4P 0















00) 0$4
4) A 0) 0

a) $ M ul 0 r-I H
4 0 4) 04 to V


0- 0 4) 0 Q) 0 0






14 9) m 0 *r 0
to 0 0 0 V
0 rH ( ZP 0) cc 0






0 4 C3 $4
ago w











O E H cM V 0


0 0
o oa

SV373U











under these conditions, or with MuIFN-gamma alone. The

slope of the dose-response curve for poly I:C was much more

shallow than that for either LPS or HKLM, which yielded

similar response curves. If sufficient triggering stimulus

was added, the trypsin-treated macrophages were fully cyto-

lytic compared to controls, indicating that the effect was

one of reduction of responsiveness to LPS or HKLM, rather

than a general cytotoxic effect on the macrophages or total

loss of responsiveness. The most striking observation,

however, was that, while the triggering responses to LPS

and HKLM were substantially abrogated by prior trypsin

treatment, the response to poly I:C was unaffected.

To further ascertain the response of trypsin-treated

macrophages to activating agents, treated and control cells

were exposed to stimuli in the presence of 35S-methionine

and examined for induced proteins as described by MacKay

and Russell (103). It was determined in pilot experiments

that the trypsin preparations used in these experiments

could trigger macrophages in the presence of MuIFN-gamma,

and that this activity was inhibitable by polymyxin B (data

not shown); this indicated that the trypsin was contami-

nated with endotoxin. To avoid the complications of

endotoxin-induced proteins by this contaminating endotoxin,

C3H/HeJ (endotoxin-hyporesponsive) macrophages were used in

these experiments. These mice are responsive to HKLM as a

triggering agent. Medium- and trypsin-treated macrophages











were exposed to 10 U/ml MuIFN-gamma in the presence of

various concentrations of HKLM in methionine-free HMEM/10%

containing either 35S-methionine (for lysate analysis) or

an equal volume of PBS (for cytotoxicity determination).

Lysates from populations of macrophages that expressed

large differences in tumoricidal activity at the same doses

of stimuli (70 and 17% for control and trypsin-treated

macrophages, respectively) were analyzed by two-dimensional

gel electrophoresis. Fluorograms of these gels are

depicted in Figure 3-7. The expression of p47b, a protein

previously shown to be associated with macrophage priming

(103), was intact in both control (B) and trypsin-treated

(C) macrophages. The expression of p71/73, a pair of

coordinately regulated proteins previously shown to be

induced by triggering stimuli, was also unaffected, despite

large differences in tumoricidal activity. The expression

of other proteins associated with induction by these

stimuli was also unchanged in these macrophages. Neither

p47b nor p71/73 was induced in C3H/HeJ macrophages pre-

treated with trypsin and then incubated in medium alone

(data not shown).


Discussion

Macrophage activation for tumor cell killing usually

occurs in a stepwise fashion as the result of exposure to

multiple stimuli. The two stages under investigation here,












0

(ae V1
r 4.) Ol 4
1 r-i V a )
rZ rI 00
1 H H
0I 04 H- ) *
SI o & H a>
*O 0 10^X i
c 0 a .0 P 0. *0
0 4- B M () 0
0 CO ) 4 0 M -I.-
to Q).i 4- oo a1 a) 4.)c
I 4* 4O1U urn l
V -r4 0 44 CO *0 WU )
r: 0 tv 4)OH 4>
- i -40 0 XO0r-4

) 0 r 0 a 0 r
V 0 ^ a ) 04.>) 0
C 4 40) 4 000 M 1
0. 0II. AC4A1
O MO 10 0ft-
o ) t4C 0 O 0 0)V 0.
O4 $4W )U4) ( 1~0 Oo
0 0 ao 3 em
o k k ra~ 10 -H.0Q
04 a4) ) -A a 0'
0 k P U >1 1 0 a
-H (DO4 0 V) 0> 0 4 a


O4-404 -HO)44000r
U 0 040 & 0 00)
3 in 04 -H u 01 aO


am % 0 N -H 00o4a
) .0V4) 0 0 U ko
.94r) 4U.)r-440.) r00


0 k400 (D 00 k
A3 0 () W >14.) > k 0 0
4J )4. > U >1T -l l 0

tU U4.) 0 6i
0 M Q 144 r3 U C 0.01
0 :3 0o ) Da (4.) 4) V
U c0 r-4 0# U I 0 k)
41- YO r4 (D O 0m 3 .3 (04.)
C U )-CA *) V 0

0 V a 0 -H XN 0 k
4) c ) %a -HH 01
e OO a! am>lP
I ke)e'W kiV00 0
St0 4 r- N 0) (0 a
$4 1 PO k (a W-1 >HOr-4 rl 44 W
4)10 0* *Ht -40 ml :9
0 .0 0 r. a k 0 >i

0 4) 0 0 0
00 0@ 2: 0 0@4 4 v
ha Oo meh j d

a u *r- AH0oA. H+
I k a t *r HO<
T r 00am 41 0 m Wo +
4 0 B0 I- < )D k 4
41 R ) 0 0 m OC 0 4. I
4Ok 0 ( 0 k O a
0 k 0 k4) 00 x0A 0H
ft CMiXS-OMB* UH
(a 0 3P k 4.) 3 0i 44 r
CA 0. 4 ) *- ,CA
40 .0 04.) 0o o *- 0
UM Tl(U^lSO-POU^ fCl
*MU ao -4C ra 0O Am ..
4) -4 t3 0 9Wk 0)H 0' 4
64.4 -H 4) t0 4) U $4 0 H
0)1 --. 4 4) F.o V 0 :D
4- m I 0 i A00 k-.
01 f : f 0 0 4J 4 0
*I U*rINO~k00k H
0. 4 004 1 .9




0>Tl
Ps 41
fn V >t 0M 0004 0 4e *
$4 r-4 ro. kre n (1 r
4. 44J4)4








!P.. '~'


) *
1 '[.'


SW
0 I
S
Do, ,0,
I m ~e *


I

* 4~j4


oew

a1


AO a" ,











priming and triggering, represent the final steps in the

modulation of a complex function (tumoricidal activity).

Understanding the subcellular events required for the full

induction of activation is of primary importance in

revealing the mechanisms behind the regulation of

macrophage tumoricidal function. The results reported here

confirm the requirement for a cell-surface receptor for

gamma interferon for priming for cytolytic activity in

macrophages, and implicate the requirement for other

protease-labile structures in the triggering response as

well. These results also suggest the existence of at least

two different pathways for triggering of primed

macrophages, and separate the induction of triggering

agent-induced proteins from the functionally important

triggering event.

The existence of specific cell-surface receptors for

IFN-gamma on macrophages has been documented for both mouse

(4,27,89) and human (26,50,128,143) cells. There have been

few data, however, to directly demonstrate the functional

requirement for these receptors for macrophage priming,

with the exception of the effects of blocking antibodies

(151,163). The data shown here demonstrate that the MuIFN-

gamma receptor on macrophages is susceptible to proteolysis

by a broad spectrum protease (Pronase), and that the

digestion of the receptor is accompanied by a loss of

macrophage responsiveness to the ligand. This confirms











that the receptor is, at least in part, a protein and the

fact that it is essential for the induction by MuIFN-gamma

of non-specific tumor cell killing by macrophages. The

binding activity for 125I-MuIFN-gamma returned fully after

removal of the protease within 4 hours; this recovery was

dependent upon de novo protein synthesis, as it did not

occur in the presence of cycloheximide (Figure 3-5). This

shows that 4 hours is the period of time required to

synthesize and express MuIFN-gamma receptors on the cell

surface. In addition, these data indicate that the

receptor is constitutively synthesized. Pronase-treated

macrophages did not recover their capacity to express

cytolytic activity in response to MuIFN-gamma and LPS for

another 4 hr, i.e., until 8 hours after removal of the

enzyme (Figure 3-4). This is consistent with the fact

that, once functional receptors are available, and at

concentrations of stimuli that are fully stimulatory, it

requires approximately 4 hours to generate the cytolytic

response (Figure 3-1). In addition to the MuIFN-gamma

receptor, any other protease-labile structures required for

activation would have to be regenerated within this same

time.

In contrast to the human IFN-gamma receptor (26), the

MuIFN-gamma receptor does not appear to be susceptible to

trypsin digestion. This fact was confirmed at three

levels. First, specific binding of 125I-MuIFN-gamma was











unaffected by trypsin treatment (Figure 3-3). Second, the

response of trypsin-treated macrophages to MuIFN-gamma for

induction of the interferon-induced protein, p47b remained

intact (Figure 3-7). Finally, the functional response of

these cells to MuIFN-gamma was also unaffected, as they

readily became activated if provided the proper triggering

agent (Figure 3-6; poly I:C panel).

A novel finding revealed by these experiments is the

distinction of triggering responses between LPS/HKLM and

poly I:C. Earlier studies by Taramelli and Varesio (183)

suggested that such a difference exists; they found that

resident peritoneal macrophages were refractory to LPS but

not to poly I:C activation. The resistance of the poly I:C

response to trypsin treatment (Figure 3-6) indicates that

this triggering agent probably acts via a pathway different

from that of LPS or HKLM. These data also suggest that LPS

and HKLM act by a similar, if not identical, mechanism that

is susceptible to alteration by trypsinization. Trypsin

treatment affected the HKLM response equally in both

C3H/HeN and C3H/HeJ macrophages. If LPS and HKLM interact

with the same trypsin-sensitive structure, this would

suggest that the LPS hyporesponsiveness in HeJ macrophages

is due to some subtle alteration of this structure, perhaps

resulting in a lower affinity for the LPS molecule.

Presumably, the post-receptor (for HKLM) mechanisms are

intact in these cells. This is an important finding with











respect to understanding the lack of LPS-responsiveness of

C3H/HeJ macrophages. Finally, these results have separated

the induction of expression of LPS/HKLM-induced proteins

(specifically p71/73) from the actual triggering event

(Figure 3-7). This finding is compatible with the fact

that triggering of primed macrophages for induction of

cytolytic activity is defined by a rapid (<15 minutes)

response that would not be expected to depend upon the

synthesis of new proteins. Trypsin affected the triggering

response without affecting the LPS/HKLM-induced proteins

p71/73, as well as others that have previously been linked

to LPS stimulation of macrophages (103). There are at

least two possible explanations for this result. 1)

Trypsin is digesting a cell-surface protein(s) required for

the triggering response, but not for the induction of

p71/73 and other LPS-induced proteins. The shift in dose-

responses to triggering stimuli rather than a complete loss

of response suggests that trypsin may not be completely

digesting these structures, but cleaving them so as to

reduce their affinity for the ligands (LPS, HKLM). 2)

Trypsin acts in a pharmacological manner, providing or

inducing a signal that negatively regulates triggering by

LPS or HKLM. The enhancement of C3H/HeJ B-cell

responsiveness by trypsin treatment has led Kuus et al.

(86) to speculate that LPS activates a membrane-associated

protease that results in signal transduction for B-cell











proliferation. In either case, these data suggest the

importance of a trypsin-labile substrate that is necessary

for LPS responses, particularly triggering for tumoricidal

activity.

In conclusion, the use of cell surface proteolysis has

revealed several observations related to the induction of

macrophage activation. 1) The receptor for MuIFN-gamma is

resistant to functional alteration by trypsin, in contrast

to the human receptor (26). 2) The MuIFN-gamma receptor,

in addition to other trypsin-labile structures, is required

for the induction of activation by MuIFN-gamma in the

presence of LPS or HKLM. 3) Lipopolysaccaride or HKLM and

poly I:C appear to act via two distinct pathways for

triggering of primed macrophages. 4) The triggering

response has been distinguished from the induction of

proteins synthesized in the presence of these same stimuli.
















CHAPTER IV

SOLUBILIZATION AND QUANTIFICATION OF THE RECEPTOR
FOR MURINE GAMMA INTERFERON


Introduction

Gamma interferon, a product of stimulated T

lymphocytes, induces multiple alterations in the function

and characteristics of mononuclear phagocytes. These

result in an overall enhancement of the host defense

functions performed by these cells (reviewed in 1,161,192,

203). Among the more complex changes regulated by this

lymphokine is the development of tumoricidal activity

(133,164). Characterization of a functional receptor for

gamma interferon on macrophages is the first step toward

understanding how it induces activation for tumor cell

killing. To isolate and definitively characterize the

binding site, the receptor must first be solubilized in an

active form, and a means found to quantify binding activity

during subsequent purification steps. The work to be

described here has successfully addressed these needs,

using as starting material the mouse myelomonocytic cell

line, WEHI-3. Gamma interferon binding activity was

detected in detergent-solubilized WEHI-3 membrane lysates

by a liposome-based precipitation assay. Use of this












soluble receptor assay confirmed the expectation that IFN-

gamma binding activity could be significantly enriched by

affinity chromatography on IFN-gamma-Sepharose.



Materials and Methods

Cell Culture and membrane preparation. WEHI-3 cells

were passed daily in 1500-ml spinner flasks in Dulbecco's

modified Eagle's medium (DMEM) supplemented with 2 mg/ml

sodium bicarbonate, 2 mM glutamine (all from Flow Laborato-

ries, McLean, VA), 100 U/ml penicillin G potassium, 100

Ug/ml streptomycin (both from Pfizer, Inc., New York, NY),

15 mM HEPES (Sigma Chemical Co., St. Louis, MO), and 10%

fetal bovine serum (FBS; Sterile Systems, Inc., Logan, UT).

RAW-264, PU-5, and P388D1 macrophage cell lines were grown

in HEPES-buffered modified Eagle's medium supplemented as

above. J774.1 cells were cultured in DMEM/107. All of

these lines were passed three times weekly in 100 mm square

Petri dishes.

Membrane preparation. Membranes were prepared accord-

ing to the method of Thom, et al. (184), with slight

modifications. WEHI-3 cells were harvested from spinner

culture and washed three times with borate-buffered saline.

Membranes were extracted by adding cells dropwise to 20 mM

boric acid containing 0.2 mM EDTA. Crude membranes were

collected after passing the lysate through glass wool by

centrifugation at 12,000 x g for 30 min at 40C. Pellets




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 ETXJQ5T7U_V07WRN INGEST_TIME 2012-02-20T21:22:05Z PACKAGE AA00009094_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES