Chapter 15 Crowd Control: Tumor Immunology and Immunotherapy It
is by no means inconceivable that small accumulations oftumour
cells may develop and, because of their possession of new antigenic
potentialities, provoke an effective immunological reaction with
regression of the tumour and no clinical hint of its existence.
Macfarlane Burnet, immunologist, 1957 Throughout this text, we have
studied various defenses that the body erects against the
appearance of cancerous growths. Many of these defenses are
inherent in cells, more specifically in their hard-wired regulatory
circuitry. The most obvious of these are the controls imposed on
cells by the apoptotic machinery, which is poised to trigger the
death of cells that are misbehaving or suffering certain types of
damage or physiologic stress. The pRb circuit and the DNA repair
apparatus are similarly configured to frustrate the designs of
incipient cancer cells. The organization of tissues also places
constraints on how incipient cancer cells can proliferate. For
example, normal epithelial cells that lose their tethering to the
basement membrane activate the form of apoptosis that is called
anoikis. This mechanism limits the ability of epithelial cells to
stray from their normal locations within tissues and grow in
ectopic (i.e., abnormal) sites. At the same time, the special
status afforded to stem cells and their genomes (Section 12.3) also
reduces the probability of mutant cancer cells' gaining a foothold
within a tissue. 655 Chapter 15: Crowd Control: Tumor Immunology
and Immunotherapy Beyond these cell- and tissue-specific
mechanisms, mammals may have another line of defense-the immune
system. The immune system is highly effective in detecting and
eliminating foreign infectious agents, including viruses, bacteria,
and fungi, from our tissues. One of the major questions in cancer
research over the last half century has been whether the immune
system can also recognize cancer cells as foreigners and proceed to
kill them. Actually, evidence is rapidly accumulating that the
immune system does indeed contribute to the body's multilayered
defenses against tumors. The difficulties associated with
establishing this type of anti-cancer defense are apparent from the
outset: the immune system is organized to recognize and eliminate
foreign agents from the body while leaving the body's own tissues
unmolested. Cancer cells, however, are native to the body and are,
in many respects, indistinguishable from the body's normal cells.
How can cancer cells be recognized by the immune system as being
different and, therefore, appropriate targets of immune-mediated
killing? We will wrestle with this problem and its ramifications
repeatedly throughout this chapter. The field of tumor immunology,
more than any other area of cancer research, remains in great flux,
with basic concepts still a matter of great debate. Consequently,
in this chapter, you will find many observations and conclusions to
be much more tentatively stated than elsewhere in this book and
subject, no doubt, to future revision. Still, this is an area of
cancer biology that is well worth our time and study, since it
holds great promise for new insights into cancer pathogenesis and
new ways of treating hwnan tumors. Research conducted on mammals
over the past three decades has.revealed an immune system of great
complexity and subtlety. Before we enter into discussions of its
anti-tumor functions, we need to take an excursion into the
workings of the general immune system. An understanding of its
mechanisms of action, at least in outline, is a prerequisite for
engaging the three major questions that will occupy us in this
chapter. First, what specific molecular and cellular mechanisms
enable the immune system to recognize and attack incipient cancer
cells? Second, do these immune mechanisms represent effective
defenses that prevent the appearance oftwnors? Third, how can the
immune system be mobilized by oncologists to attack tumors once
they have formed? (An introduction to immunology will occupy our
attention in Sections 15.1 through 15.6; an overview will be
provided in Figure 15.14.) 15.1 The immune system functions in
complex ways to destroy foreign invaders and abnormal cells in the
body's tissues The mammalian immune system launches several types
of attack against foreign infectious agents and the body's own
cells that happen to be infected ""ith such agents. It identifies
its targets by recognizing specific molecular entitiesantigens-that
are made by these agents. Having done so, the immune system
undertakes to neutralize or destroy the infectious particles
(bacterial and fungal cells, virus particles) , as well as infected
cells displaying these antigens. To the extent that the immune
system also functions to ward off cancer, one assumes that it
exploits many of the same mechanisms that it uses to eliminate
foreign infectious agents. The most familiar of the immunological
defense strategies involves the humoral immune response-the arm of
the immune system that generates soluble antibody molecules capable
of specifically recognizing and binding antigens (Figure 15.1).
Thus, a virus particle or bacterium displaying antigens on its
surface may rapidly become coated with antibody molecules, which
may result in the neutralization of these pathogens (Figure 15.2).
Similarly, an infected cell 656 (C) '.1-- light Function of the
humoral immune response lig ht (L) chain (B) constant region di su
lfide bonds (A) Figure 15.1 Structure of antibody molecules and
their binding to antigens The most abundant antibody molecule in
the plasma is the immunoglobulin y (lgG) molecule. (A) X-ray
crystallography of an IgG molecule reveals the symmetry that allows
the two antigen-binding domains (top left, top right) to bind two
antigen molecules Simultaneously. (B) IgG molecules are divided
into two functional regi ons. One region is designed to recognize
and bind antigen molecules. Because the IgG molecules in plasma can
recognize an essentially unlimited number of antigens, IgG
molecules have a comparable diversity of structures in their
antigen-binding portions, which are call ed their variable domains
(red), to recognize this diversity The remainder of the IgG
molecule is termed its constant region (blue) and is invariant am
ong all IgG molecules of a given subclass, e.g, all IgG 1
molecules. An IgG molecule as a whole is a heterotetramer composed
of two light (L) chains and two heavy (H) chains. Two separate
antigen-recognizing and binding pockets are displayed (top left,
top right), each composed of an H- and an L-chain N-terminal domain
(concave shapes). (C) The detailed struct ure of an
antigen-antibody complex is seen here in this space-filling
molecular model in which the antigen-binding domains of the heavy
chain (purple) and light chain (yel/ow) are seen to contact the
antigenic molecule, in this case the chicken egg-white lysozyme
molecule (light blue). Only parts of the variable regions of the
heavy and light chains are shown here. (A glutamine residue, red,
is indicated that is important for the hydrogen bonding of the
antigen to the antibody molecule.) (From CA Janeway Jr. et ai,
Immunobiology, 6th ed . New York Garland Science, 2005 ) may
display on its surface the antigens made by the agents that have
infected it and its surface may become coated with antibodies that
recognize and bind these antigens. Once a mammalian cell or an
infectious agent is coated by antibody molecules, it may be
recognized, engulfed, and destroyed by phagocytic t- A ~ l'--J;). ~
; . . - < ~ lJ).. ant igen (chicken egg-White lysozyme) heavy
(H) chain (L) chain Figure 15.2 Neutralization by antibody
molecules (A) Virus particles (red) can become coated by antibody
molecules (blue) developed by the immune system of an infected
host. This coating neutralizes (inactivates) the infecti vity of
the particles by blocking their adsorption to host cells . (B)
Similarly, a bacterium displaying certain surface antigens (red)
can also be prevented from adhering to host cells (A) antibody
prevents viral adsorption (B) antibody prevents bacterial adherence
by bound antibody molecules. 657 Chapter 15: Crowd Control: Tumor
Immunology and Immunotherapy (A) bacterium Fc receptors ~ -o
macrophage - -{ ~ lysosome (B) Fc receptors I(\ I / NK cell
activated~ ~ ~ ~ ~ ~ - -NK cell o 0 0 targeted mammalian cell (e)
Figure 15.3 Coating of pathogens by antibody molecules and their
elimination by effector cells of the immune system The coating of
viruses, bacteria, and mammalian cells by antibody molecules is
often the prelude to their being phagocytosed (engulfed) or
destroyed by cytotoxic cell s of the immune system. (A) The coating
of a bacterium (red) by antibody molecules (yellow) may provoke a
macrophage to use specialized receptors on its surface, termed Fc
receptors (green), to recognize and bind the constant regions of
the antibody molecules (which are not involved in antigen
recognition; see Figure 15.1). This often result s in the
phagocytosis of the antibodycoated bacterium and its eventual
destruction in Iysoso mes within the cytoplasm of the macrophage.
(B) A mammalian cell (gray) becomes coated by antibody molecules
(blue) that recognize and bind antigens (red) on its surface. A
type of lymphocyte termed a natural killer (NK) cell then uses its
Fc cell surface receptors (green) to bind the constant regions of
the coating antibody molecules. This binding results in acti vation
of the N K cell , which proceeds to destroy the targeted cell,
using cytotoxic granules (purple dots), whose contents it
introduces into the targeted cell, to do so. (C) Sheep red blood
cells were treated with an antibody that recognizes an antigen
displayed on their surface. As seen in this scanning electron
micrograph, a large number of them have become adsorbed to a
macrophage via the Fc receptors on the surface of the latter. (A
and B, from CA Janeway Jr. et ai, Immunobiology, 6th ed. New York
Garland Science, 2005; C, from J. Sw anson, University of
Michigan.) cells, such as macrophages, or killed by cytotoxic
cells, such as natural killer (NK) cells (Figure 15.3).
Importantly, these immune cells do not, on their own, have the
ability to recognize specific foreign antigens. Instead, the
antibody molecules that have bound to antigens on the surfaces of
viruses, bacteria, or mammalian cells alert these immune cells to
the presence of targets that should be destroyed. The other arm of
the immune system involves the cellular immune response. This
response is mounted when specialized cytotoxic cells are developed
by the immune system that can, on their own, recognize and directly
attack other cells displaying certain antigens on their surface. In
this case, soluble antibodies are not required as intermediaries to
recognize antigens displayed by targeted cells, since cytotoxic
cells of the T-lymphocyte lineage (CTLs) have developed their own
antigen-recognizing machinery in the form of T-cell receptors
(TCRs), which they use to identify cells bearing particular
antigens; such cells are then targeted for destruction by the
cytotoxic T lymphocytes (Figure 15.4). 658 Antigen presentation and
the immune response We can also depict the immune system in another
dimension: many of the responses of the immune system to an
infectious agent (e.g., a specific strain of virus) and its
antigens depend on a previous encounter with this agent. The immune
system has been "educated" through the initial encounter to
recognize certain antigens displayed by this agent and to mount a
vigorous counterattack against it in the event that it encounters
this agent a second time; this represents the adaptive immune
response. At the same time, other cellular components of the immune
system are naturally endowed with the ability to recognize certain
infectious agents or abnormal cells and thus do not require prior
exposure and education; this inborn ability is termed the innate
immune response. For example, the natural killer (NK) cells cited
above have the ability to recognize specific cell-surface molecules
displayed by aberrant cells, even without having encountered such
cells previously. 15.2 The adaptive immune response leads to
antibody production Adaptive immune responses begin when infectious
particles or abnormal cells are engulfed by specialized phagocytic
cells of the immune system, notably macrophages and dendritic cells
(DCs) (Figure 15.5). Having ingested these objects or fragments
thereof, the phagocytic cells are then charged with the task Figure
15.4 Cytotoxic T lymphocytesof presenting the ingested contents to
other cellular components of the immune The cellular arm of the
immune response system. More specifically, these cells must inform
the immune system of the set results in the formation of cytotoxic
cells, of antigens that were associated with the particles that
they have ingested. This such as cytotoxic T cells (T c's, CTLs)
that presentation of ingested antigens by phagocytic cells takes
place in the lymph are able to recognize and kill other cells
nodes, to which these cells migrate following uptake of antigen.
(As discussed in displaying certain antigens on their detail below,
the antigens are presented in the lymph nodes to various types of
surface. (A) CTLs develop antibody-like T cells.) molecules on
their surface termed T-cell receptors (TCRs). A di verse array of
TCRs In order to educate the immune system, these
antigen-presenting cells (APCs) are developed during the
development first digest the particles that they have phagocytosed
(i.e., ingested outright) or of the immune system, paralleling the
endocytosed (i.e., bound via cell surface receptors and then
internalized). This development of a diverse repertoire of soluble
antibodies. Each CTL displays a digestion, which is carried out in
specialized cytoplasmic vesicles, snips internalparticular
antigen-recognizing TCR.ized proteins into small oligopeptides that
are 18-22 amino acid residues long. (B) Seen here is a CTL (upper
right, The oligopeptides are then loaded onto the major
histocompatibility complex ist panel) that has already used its
TCR(MHC) class II molecules as the latter are making their way to
the surface of APCs to recognize and bind to a target cell
(diagonally below it to the left). The cytotoxic granules w ithin
this CTL (red spot) begin over a period of minutes to (A) T-cell
receptors /! migrate toward the point of contact between the killer
and its victim. By 40 minutes, the contents of these granules (such
as granzymes; see Section 9.14) have been introduced into t he
target cell, which has already advar.ced into apoptosis and begun
to disintegrate. (From C.A. Janeway Jr. et ai, Immunobiology, 6th
ed . cytotoxic T cells (CTLs) New York: Garland Science, 2005)
target cell CTL 'j,' ... f '/',,' . ', ' ,.,'. .' /I '..:' .. ,
...., e"" "'--\'," . .. "/ ,' . . # l ; I .. \. "/-, J \ ",. 1..,
.... '...' " .'}. r'; . , .... .I.' . .. f - l. "', - .-......:-..
... ' '" .. . .' . . "."\"'''''''' .:, ,-'0 ........,' \.... -. , .
\ 11....t:....;to:'!I'. time =0 after 1 minute after 4 minutes
after 40 minutes 659 Chapter 15: Crowd Control: Tumor Immunology
and Immunotherapy (A) antigens ~ .' . .," -antigen uptake by
Langerhans cells leave Langerhans cells enter Langerhans cells in
the skin and enter the the lymph node to become the skin lymphatic
system dendritic cells expressing B7 (B) B7-positive dendritic
cells stimulate naive T cells Figure 15.5 Antigen presentation by
dendritic cells The immune syste m becomes aware of infectious
agents and their antigens largely through the acti ons of ant
igen-presenting cells (APCs), notably dendritic cells. (A) Here we
see a drawing of specialized phagocytic cells (i.e., Langerhans
cells, yellow) residing in the skin, which take up antigens (red
dots) by phagocytosi s and then migrate to the lymph nodes (light
blue), w here they mature into dendriti c cells. In the lymph
nodes, these cells confront T cells (dark blue circles), to which
they present antigens; this results in the functional activation of
the T cells and the subsequent mounting of a specific immune
response against cells and viruses that display these antigens. (B)
The dendritic cells take their name from their mUltiple arms
extending out from the cell body. (From CA. Janeway Jr. et aI.,
Immunobiology, 6th ed. New York Garland Science, 2005. ) (Figure
15.6) . More specifically, the oligopeptide fragments become
attached to the specialized antigen-presenting domains (Figure
15.7A) of MHC class II molecules. (In humans, the MHC molecules are
often termed HLA, or human leukocyte antigen, molecules, but we
will use the more generic term, MHC, throughout this chapter to
refer to both human and murine molecules of this type.) The class
II MHC molecules function much like a street hawker's hands used
for displaying wares to passers-by. In this case, the wares are
oligo peptide antigens captured by the APCs and the intended
customers are other cells of the immune system, specifically a
class of lymphocytes termed hel per T cells (TH cells) , often
called CD4+ cells to reflect a specific cell surface antigen that
they display ~ reticulum MHC class 1\ Figure 15.6 Antigen
processing by antigen-presenting cells in the endoplasmic
reticulum, which then move to the cell surface, After phagocytes,
notably dendritic cells and macrophages, have allowing the MHC
class II molecules to display the oligopeptide internalized
potential antigenic particles (red oblongs), these are fragments on
their surface and present the oligopeptide fragments fragmented
into oligopeptides (red dots) by proteolysis. The to T cells in the
lymph nodes. (From CA Janeway Jr. et al., resulting oligopeptides
are then loaded onto MHC class II molecules Immunobiology, 6th ed.
New York: Garland Science, 2005.) 660 Antigen presentation and the
immune response oligopeptide antigen oligopeptide antigen Figure
15.7 Antigen presentation by MHC molecules (A) The structure of the
antigen-presenting groove of an MHC class II molecule is shown here
as determined by X-ray crystallography. The oligopeptide antigen
(stick figure, in color) that is bound via hydrogen bonds to the
"palm " of the MHC molecule (ribbon diagram) is shown w ith its
N-terminus to the left and C-terminus to the right. The oligopepti
de antigen together wit h the nearby amino acid residues of the MHC
molecu le form the molecular structure that is recognized by other
immune cells, which may, for example, use T-cell receptors to do
so. (B) A very similar arrangement characterizes the structure of
the antigen-presenting domain of MHC class I molecules. (From C.A.
Janeway Jr. et al., Immunobiology, 6th ed. New York: Garland
Science, 2005.) (Figure 15.8). Because macro phages and dendritic
cells are specialized to use their MHC class JI molecules to
present antigens scavenged from their environment, immunologists
sometimes call them "professional" APCs, to distinguish them from
other types of cells that are not specialized for this type of
antigen presentation. Note that it is the combined molecular
structures formed by the class II ectodomains (t he "hands") and
their bound oligopeptide antigens (the "wares") that are presented
to helper T (TR) cells (see Figure 15.7). Antigen presentation to
certain helper T cells provokes the latter to activate, in turn,
the B cells that can manufacture immunoglobulin (antibody)
molecules that specifically recognize and bind the particular
antigen (see Figure 15.8). The subsequent maturation of these B
cells yields a population of cells (called plasma cells) that
actively secrete this particular antibody species into the
circulation, that is, antibody molecules that are specialized to
recognize and bind the particular antigen that originally triggered
this series of responses. (Dendritic cells, once again functioning
as APCs, can also activate cytotoxic T cells-not indicated in
Figure 15.8.) This system works well when confronting infectious
agents such as virus particles, bacteria, and fungi in the
extracellular spaces. Thus, these infectious agents can be
internalized by the professional antigen-presenting cells, and the
peptides deriving from the ingested agents can be presented again
to the outside world. The antibody molecules that are eventually
formed by B cells and their descendants as a result of this antigen
presentation can recognize and bind the infectious particles and
thereby neutralize them (see Figure 15.2). By the same token, we
can imagine that cancer cells displaying certain distinctive
antigenic proteins on their surfaces might also provoke an antibody
response by the immune system and become coated by antibody
molecules bound to these cell surface antigenic molecules. 661
Chapter 15: Crowd Control: Tumor Immunology and Immunotherapy (A)
dendritic cell productive interaction between APC and T H cell TH
cell antibody molecules activatio'n differentiation TCR secreted (]
n V !J B-cell plasma cell activation (B) TH cells Figure 15.8
Immunocyte encounters within lymph nodes Dendritic cells interact
directly wit h helper T cells and present antigen to them in the
lymph nodes. (A) Dendritic cells engulf, process, and present
antigenic oligopeptide fragments (red dots; see Figure 156) on
their surface to T cells in the lymph nodes, using their MHC class
II molecules (gray) to do so. Here, a dendritic cell meets a number
of T cells (above), known hereafter as helper T (T ~ cell s. Each
of them displays its own distinCt T-ce" receptor (TCR; green) on
its surface. However, in the f irst three encounters, none of the T
H cells' T-ce" receptors recognizes and binds the antigen being
presented by the MHC II molecule of the dendritic cell.
Nonetheless, on occasion, the dendritic cell succeeds in finding a
T H cell whose T-cell receptor does indeed recogni ze the
oligopeptide antigen being presented by the dendritic cell's MHC
class II molecules (belovv) This causes the T H to become
activated; the T H cell leaves the dendritic cell and proceeds to
search for B cells that also display on their surface the same
antigen in t he cont ext of M HC II. When and if the T H finds such
a B cell (light yellow, 2nd diagram from righ t), it activates the
Bee", which proliferates and, having differenti ated into a plasma
cell (light brown), begins to release antibody molecules that are
capable of recognizing thi s oligopeptide ant igen. (B) Multiphoton
microscopy reveals the capsule of a mouse lymph node (blue) and a
number of recently arrived, dye-labeled dendritic cells (red dots)
as well as dye- labeled T cells (green dots) to which antigen wi ll
be presented by the dendritic cells. The two cell types are largel
y segregated from one another within the lymph node, and their
mechanisms of trafficking and interact ion within the lymph node
remain poorly understood. The T H cells have arrived in the lymph
node from the venous circulation and have extravasated via
diapedesis (Sidebar 14.3) in order to take up residence in the
node. (B, from TR. Mempel, S.E. Henrickson and U.H. Von Andrian,
Nature 427: 154-1 59, 2004) 662 --Cytotoxic cells and the immune
response (A) antibody-antigen \ cell surface complexes channel
inserted in plasma membrane (B) The antibodies coating a cell or
infectious agent may elicit an alternative type of immune attack: a
set of proteins in the plasma, termed complement, will recognize
the constant regions of antibody molecules tethered to the surface
of a cell (including bacterial, fungal, and mammalian cells). bind
to these antibody molecules, and proceed to punch holes in the
adjacent plasma membrane, thereby killing the cell (Figure 15.9).
This series of steps leading to adaptive humoral responses tells us
something important about the molecular structure of the antigens
that are immunogenic, that is, that elicit immune responses: they
are not intact proteins, but instead are oligo peptide fragments
derived from the cleavage of much larger proteins (see Figures 15.6
and 15.7). (The major exceptions to this generality are certain
complex carbohydrate chains and linked side chains that may, under
some circumstances, also be immunogenic.) 15.3 Another adaptive
immune response leads to the formation of cytotoxic cells The type
of immunologic response described above fails to deal effectively
with infectious agents that have entered into cells and are
therefore shielded by the plasma membrane from scrutiny. Similarly,
in the case of cancer cells, the humoral response system will fail
to recognize aberrant cellular proteins that are hiding deep within
these cells. In principle, such shielding should create a serious
problem for the immune system, which needs to monitor what is going
on inside cells in addition to its task of monitoring the contents
of the extracellular spaces and the surfaces of cells. The problem
is solved by an antigen-presenting mechanism that echoes the one
used by the professional antigen-presenting cells (APCs) described
above. Actually, this other antigen-presenting mechanism is the
more widespread of the two, since it is used by the great majority
of cell types throughout the body. It works like this (Figure
15.10): rather than being used for their normally designated
functions, a portion of the proteins synthesized within cells (by
some Figure 15.9 Complement-mediated killing (A) Antigen-antibody
complexes (red spheres) formed by the binding of antibody molecules
to cell surface antigens (left) can attract complement proteins
present in the plasma (ye//ow, green, purple) and induce them to
form complexes that lead, through a series of steps, to the
formation by other complement proteins of channel s in the plasma
membrane of the cell (right) at a site adjacent to w here the
antigen-antibody complexes initially formed . (B) The resulting
channels, seen here in this electron micrograph, destroy the
integrity of the barrier functions of the plasma membrane and lead
rapidly to cell death. (From CA Janeway Jr. et al., Immunobiology,
6th ed. New York Garland Science, 2005) 663 Chapter 15: Crowd
Control: Tumor Immunology and Immunotherapy (A) to cell surface,
.cytosol ..... membrane (B)f Og re 15.10 Display of intracellular
antigens by MHC class I molecules vesicle _____ MHC class I
endoplasmic reticu lum oroteasome .. plasma -q ..I peptide
fragments :.. s: all cell types throughout the body, including
cancer cells, routinely divert G OO L on of their recentl y synt
hesized proteins to the antigen-presenting acmnery. (A) Some of the
recently synthesized proteins in the cytosol are .l1ed t o
proteasomes (purple, yellow), in which they are broken down into
gopeptides (red dots); resulting oligopeptides are then introduced
into the e1doplasmic reticulum, where they may encounter MHC class
I molecules yellow) that bind them relati vely tightl y (see Figure
15.7B). Thi s will cause the mul ti-protein complexes to move via
membranous vesicles to the cell surface, w here these protein
complexes serve to display to the immune system fragments of the
proteins that are being synthesized within the cell. The overall
process of displaying these antigens is similar to that undertaken
by MHC class II molecules (Figure 15.6); however, MHC class II
antigen presentation is the speciality of "professional
antigen-presenting cells", such as macrophages, dendritic cells,
and B cells, while MHC class I presentation is undertaken routinel
y by almost all cell t ypes in the body. (B) A broad spectrum of
oligopeptide fragments deri ving from v MHC class I a large number
of cellular proteins (here represented as four distinct protein
species) are displayed simultaneously by cells using their MHC
class I proteins (A, from CA Janeway Jr. et ai, Immunobiology, 6th
ed. New York Garland Science, 2005.) accounts, as much as one-third
in certain cells) is routinely diverted to specialized proteasomes.
There these proteins are cleaved into oligopeptides. These cleavage
products, of 8 to 11 amino acid residues in length, are then
attached to MHC molecules en route to the cell surface and
displayed on the outside of cells by the other major class of
antigen-presenting molecules-the MHC class I molecules (see Figure
15.7B). Included among the intracellular peptides displayed by the
MHC class I molecules are those synthesized normally by cells as
well as those made by foreign infectious agents within the cell,
such as viruses and bacteria. This external presentation of
internal antigens occurs routinely and continuously, whether or not
foreign proteins happen to be present within a cell. The display by
a cell of certain oligopeptide antigens on its surface (via its MHC
class I molecules) may attract the attention of cytotoxic T cells
(Tc's, also called cytotoxicT lymphocytes, CTLs, or CD8+ cells),
which proceed to kill this cell (see Figure 15.4). The origins of
this killing can be traced back to the actions of helper T cells.
Recall that some helper T cells are able to activate the humoral
immune response by interacting with and stimulating
antibody-producing B cells (see o - cellular proteins . . -. , .. .
. .' . ..". .. ..o .ee . ,_ ..". . ...- , ........ . .. 664
Cytotoxic cells and the immune response Figure 15.11 Activation of
cytotoxic T cells by helper T cells In addition to inducing B cells
to make antibody molecules (Figure 158), helper T celis (TH) of a
second subtype (blue) can activate the precursors of cytotoxic T
celis (light red, bottom) to become active cytotoxic T cells
(termed Tc 's or CTLs, red) that can use their T-cell receptors
(TCRs) to recognize and bind antigens presented on the surfaces of
many cell types throughout the body by MHC class I molecules. This
recognition results in attack on the antigen-displaying cell (gray,
top), as shown by the micrographs of Figure 15.4B. The Tc 's often
use cytotoxic granules (black dots) containing perforin and
granzymes to kill targeted cells (Figure 15.12) Figure 15.8). Now,
we encounter a second, independent function of helper T cells: some
of them can contribute to the activation of cytotoxic T cells,
which are specialized to recognize and kill target cells displaying
the particular oligopeptide antigen that initially provoked an
immune response (Figure 15.11) . This attack on antigen-displaying
cells by cytotoxic lymphocytes represents the cellular arm of the
adaptive immune response. The capacity of helperT cells to
facilitate development of both humoral and cellular immune
responses reflects the ability of distinct subpopulations ofTH
cells to produce and release the soluble immune factors known as
cytokines: TH'S that promote hwnoral immunity (by stimulating B
cells) produce interleukin-4 OL-4), while TH'S that promote
cell-mediated immunity (by stimulating cytotoxic T cells) secrete
interferon-y (IFN-y) . Cytotoxic T cells can kill their cellular
victims through two separate mechanisms. They can expose their
intended victims to certain toxic proteins (Figure 15.12A and B) .
One of these, periorin, punches holes in the plasma membrane of a
targeted cell . These holes then enable granzymes released by the
cytotoxic cells to enter into the cytoplasm of the victim. As
described earlier (Section 9.14), once in the cytoplasm of the
targeted cell, granzymes cleave and thereby activate pro-apoptotic
caspases. The second killing mechanism, also discussed in Section
9.14, involves the Fas death receptor, which is displayed on many
cell types throughout the body. Cytotoxic T cells can present the
ligand of the Fas receptor, termed FasL, to their intended victims.
FasL then activates the Fas death receptors on the surfaces of the
targeted cells, thereby activating their extrinsic apoptotic
pathway (Figure 15.12C) . Killing by cytotoxic T cells can play an
important role in limiting the infectious spread of viruses. For
example, a recently infected cell in which a virus is actively
replicating will use its MHC class I molecules to display oligo
peptide antigens derived from cleaved viral proteins. This antigen
display will warn the immune system that abnormal proteins are
being produced deep within the cell. If the immune system is
functioning well, its cytotoxic T cells will recognize the viral
oligopeptide antigens displayed by the cell's MHC class I molecules
and kill this cell long before the virus has had a chance to
multiply and release progeny virus particles. This means that the
immune system actually uses two arms of the adaptive immune
response to limit viral infections: the cellular response is used
to kill virus-infected cells, while the humoral response is used to
neutralize virus particles that have been released into
extracellular spaces, including the circulation, by coating these
particles with antibody molecules (see Figure 15.2A). As we will
see, the anti-viral responses are important means by which the
immune system blocks the appearance of virus-induced human tumors.
antigen-displayi ng target cell MHC TH cell L activation [ active T
c maturation T c precursor 665 Chapter 15: Crowd Control: Tumor
Immunology and Immunotherapy (A) (8) (C) targeted ce lis "as FADD
death. -X domains .. - -J pro- .fl 8 ",::::'i Q lb active cClspase:
3 . o . .target cell pro- 0. . . .. caspase 3 .. .0 . f apoptosis
(0 ) (E) 60 min-15.4 The innate immune response does not require
prior sensitization Ninety-nine percent of the animal species on
the planet do not possess adaptive immune responses to protect them
from attack by pathogens. These organisms rely on innate
immunological responses for such protection. Importantly, this
ancient, widespread innate immunity system has been conserved
during the 666 Cytotoxic cells and the immune response Figure 15.12
Mechanisms of cell killing by cytotoxic lymphocytes (A) This
electron micrograph of a cytotoxic T lymphocyte (T(, CTL) reveals a
series of lytic granules in its cytoplasm (pink arrows, left
panel). When contact is made with a targeted cell (which was
initially recognized by the T-cell receptors borne by the T C),
these granules release perforin, which forms cylindrical channels
in the plasma membrane of the target (center cel/, right panel);
pro-apoptotic proteins such as granzymes (see Section 9.14), which
are also carried in these granules, are then introduced through
these channels into the cytoplasm of the targeted cell, where they
initiate the apoptotic cascade by cleaving procaspases. (B) In the
absence of a cellular target, the lytic granules (green, yellow), w
hich contain perforin and granzymes, are scattered throughout the
cytoplasm of cytotoxic T lymphocytes (Tcs; upper panel) In the
lower panel, a synapse has been formed with a targeted cell (left),
and the lytic granules have congregated at the synapse in
preparation for killing the targeted cell. (C) An alternative
mechanism of killing cells that have been targeted for destruction
depends on the display of FasL (orange) by the T c (top, pink).
FasL, which is a trimer, then engages the Fas receptor (brown)
displayed by the targeted cell (bottom cell, gray) and triggers
receptor trimerization and resulting activation of the ext rinsic
apoptot ic cascade in the targeted cell via the sequential
activation of caspases 8 and 3 (see al so Figure 9.31). (D) NK
cells are programmed to recognize and kill other cells, including
cancer cells, that do not display normal levels of MHC class I
molecules on their surface. This scanning electron micrograph (SEM)
reveals that NK cells (colorized green), one of which has spread a
portion of its cytoplasm across the surface of a human ductal
breast carcinoma cell in the initial stage of such an attack. (E)
This SEM reveal s the initial attack of an NK cell (left panel,
smaller cell) on a leukemia cell. Sixty minutes later, the NK cell
has caused extensive damage to the plasma membrane of the leukemia
cell, which has fragmented and rolled up its plasma membrane in
response t o this attack (right panel). (A, C, and E, from C A
Janeway Jr. et ai., Immunobiology, 6th ed. New York: Garland
Science, 2005; B, from R.H. Clark, J.C Stinchcombe, A. Day et ai.,
Nat. Immunol. 4:1111-1120, 2003; D, courtesy of S.C Watkins and R.
Herberman .) evolution of mammals and continues to pLaya critical
role in various immunological responses. The cellular components of
the innate immune response are able to recognize and attack foreign
particles and aberrant cells without having been "educated" through
prior exposure to these agents. Thus, these immunocytes (cells of
the immune system) "instinctively" recognize aberrant cells, such
as cancer cells, in the body's tissues and target these cells for
attack and destruction. Instead of recognizing specific antigens,
the cells mediating innate immunity recognize characteristic
molecular patterns that are present on the surfaces of infectious
agents (or transformed cells) but are not displayed by normal
cells. A major component of the innate response is the natural
killer (NK) cell. It is likely that many initial encounters of the
immune system with cancer cells are made by NK cells. As we will
discuss in greater detail later, the NK cells recognize
configurations of cell surface proteins displayed by a wide variety
of cancer cell types. Hence, NK cells are "pre-programmed" to
recognize cancer cells and to eliminate them from the body's
tissues. In addition to NK cells, yet other cellular components of
the innate immune system, including macrophages and neutrophils,
are likely to contribute to mounting innate immune responses
against cancer cells. After an NK cell has initiated the innate
immune response by recognizing and attacking a target cell (Figure
15.120 and E), it sends out cytokine signals, notably interferon-y
(IFN-y), in order to recruit yet other immune cells, including
macrophages, to the site of attack. The actions of this second wave
of 667 Chapter 15: Crowd Control: Tumor Immunology and
Immunotherapy Figure 15.13 Destruction of normal tissues by
autoimmune attack Extensive tissue damage can be wrought b an
immune system that has been pro oked to attack the body's normal
tissues. In principle, the same immune mechanisms that are at work
here can also 0 , erate to attack and destroy malignant ti ssues.
(A) A normal pancrea ic islet (i .e , islet of Langerhans) in a
mouse (left panel) is composed largely 0 ' InSuli n-secreting
pcells (light brown) ith a small number of a cells at its periphery
(dark brown). The pancreas of a mouse sufferi ng from diabetes
resulting f rom autoimmune attack on pcells is seen 0 have lost
almost all of them (right panel). (B) A normal glomerulus in the
kidney (center of left panel), which contains a complex network of
tubules, is responsib le for the filtering of plasma and the
production of urine. In the disease of systemic lupus erythematosus
(SLE), an autoimmune attack on the basement membrane located
beneath the epithelial cells of the glomerulus results in the
accumulation of antibody protein and the invasion of a variety of
inflammatory cells; together, these eventually destroy the
architecture of the glomeruli (right) and thus kidney function. (A,
from CA Janeway Jr. et ai, Immunobiology, 6th ed . New York Garland
Science, 2005; B, courtesy of A.B. Fogo. ) immunocytes will often
enable the immune system to mount more specific and ultimately more
effective responses, in particular, adaptive humoral and cellular
responses. For example, large numbers of cytotoxic T cells can be
mobilized by the adaptive immune response to efficiently kill
cancer cells. 15.5 The need to distinguish self from non-self
results in immune tolerance The immune system is finely tuned and
highly specific. Most critically, it must be able to distinguish
foreign proteins (e.g., those made by invading infectious agents)
from those proteins that are normally made by the body's own cells.
As a consequence, if the oligopeptides displayed by one of the
normal cells in the body are similar or identical to those
routinely encountered by the immune system, this cell will remain
unmolested by the various arms of the immune system-one of the
manifestations of immune tolerance. In fact, immune tolerance
represents the major puzzle of current immunological research: How
does the immune system learn to discriminate foreign proteins and
peptides from the body's normal repertoire of proteins?
Immunologists often refer to this behavior as the ability of the
immune system to discriminate between "nonself" and "self." A
variety of mechanisms operating during the development of the
immune system ensure that any T cells and B cells that happen to
recognize self-antigens are eliminated; alternatively, if such
cells escape elimination, their actions will be strongly
suppressed. Failure to delete such self-reactive or auto-reactive
lymphocytes from the large pool of lymphocytes in the body results
in the survival of immune cells that may target the body's own
normal tissues. Should such auto-reactive cells actually do so,
this breakdown of tolerance may lead to autoimmune diseases, such
as rheumatoid arthritis, ulcerative colitis, and lupus
erythematosus, in which the immune system dispatches antibodies and
cytotoxic cells to attack normal cells and tissues (Figure 15.13).
(A) pancreatic islet (8) kidney glomerulus ~ ~ ~ normal autoimmune
destruction 668 Loss of tolerance and autoimmunity Immune tolerance
raises a simple and obvious point that will dominate the
discussions that follow: How does the immune system, which is
designed to be tolerant of the body's own cells, recognize and
attack cancer cells, which are, to a great extent, very similar at
the biochemical level to cells that are normally present in the
body? And if it does undertake attacks against cancer cells,
including those transformed by tumor viruses, how might these cells
evade and thwart the attacks launched by various arms of the immune
system (see Sidebar 36 OJ? 15.6 Regulatory T cells are able to
suppress major components of the adaptive immune response Research
beginning in the 1990s has described an entirely new class ofT
cells that have come to be called regulatory T cells (Treg cells or
simply Tregs) . Indirect evidence suggesting their existence came
from the observation that in normal individuals, a significant
proportion of cytotoxic T cells (CTLs) recognize normal tissue
antigens presented by these individuals' MHC class I molecules-a
situation that should lead directly to extensive immune attack on
normal tissues and resulting autoimmune disease. However, such
attacks do not occur, apparently because of suppression of these
cells' actions by some unknown agents. The discovery ofTreg cells
seems to have largely solved this problem, since these cells are
able to block the actions of the cytotoxic T cells that are
scattered throughout our tissues. Indeed, in genetically altered
mice lacking Treg cells, lethal autoinunune disease develops; a
comparably aggressive, ultimately fatal autoimmune disease has also
been documented in humans who are unable to make Tregs. Like T
helper (TH) cells, the Tregs display the C04 antigen on their
surface. However, the Tregs are distinguished by their additional
display of the C025 surface antigen and their expression of a
transcription factor, termed FOXP3, that programs their
development. Because Tregs express antigen-specific T-cell
receptors (TCRs; see Figure 15.4), they can specifically block the
actions of those cytotoxic T lymphocytes whose TCRs recognize the
same antigens. In addition, when located in the lymph nodes, the
Tregs can prevent the activation ofTH cells by dendritic cells. It
appears that the T regS must be in close proximity with the T Hand
T c cells that they suppress, and that the release of T G F - ~ and
interleukin-10 (IL-lO) by the Tregs is often used to inhibit or
kill these other types ofT lymphocytes. Research on TIegS is still
in its infancy. However, it is possible that their behavior holds
the key to understanding the pathogenesis of a number of autoimmune
diseases. At the same time, the actions of Tregs may explain how
many types of tumor cells can thrive in the presence of large
numbers of CTLs that should, by all rights, be able to eliminate
them-a topic pursued later in this chapter. An overview of the
various components of the immune system that we have covered until
now is provided in Figure 15.14. 15.7 The immunosurveillance theory
is born and then suffers major setbacks As suggested by the
quotation at the beginning of this chapter, the notion that the
immune system is able to defend us against cancer is an old one.
Burnet's 1957 speculation about the immune system's role in
monitoring tissues for the presence of tumors, together with other
speculations made by Lewis Thomas, represented the first instance
that the notion of the immunosurveillance of cancer was clearly
articulated. 669 Chapter 15: Crowd Control: Tumor Immunology and
Immunotherapy At the time, infecting microorganisms, specifically,
bacteria, viruses, and fungi, were known to be strongly
immunogenic, in that they usually provoke an immune response that
leads to their total eradication by various arms of the immune
system. By analogy, it was plausible that the immune system
continuously monitors its tissues for the presence of cancer cells.
Having identified them-so this thinking went-the immune system
would treat these cancer cells as foreign invaders and eliminate
them long before they had a chance to proliferate and form
life-threatening tumors. Early attempts in the 1950s to test this
model were not definitive. When tumors were removed from some mice
and implanted in others, the tumors were rapidly destroyed in a way
that gave clear indication of the actions of vigorous host immune
responses. Soon it became clear, however, that this rejection had
nothing to do with the neoplastic nature of the tumor cells.
Instead, their elimination was a consequence of what came to be
called allograft rejection. Thus, cells and tissues from one strain
of mice are invariably recognized as being foreign when implanted
in mice of a second strain. This is a consequence of the fact that
the humoral cellular immunity specific ___ __ ________ _____ __ _
__ nonspecific(adaptive) " ..- (innate) , proteinsantigen cells,-
controllers -, effectors present!ng cellscell surface antibody ,
binds antigen regulatory T NKcells, I complement(Treg) cytotoxic
dendritic celis, : .. .. __-----,,.--A____-----i macrophages . I
proliferation -st imul ate stimulate killing of targetplasma cells
,cells 1 t t innate Fc recognition receptors secrete antibodies - -
- -- -- - - - - - - -- --- - ----: -- - -- - - - - - - -- .... - --
- -- - - - -- - - - - - - -- -- - -1- --- - -- -'j! ._-----------
.. -----_._--------------- ---- ... --_._. __ ..... -.': immunity
neutralize pathogens killing of target cells __ Figure 15.14
Overview of the humoral and cellular arms of th e immune system The
humoral immune system (left) is"driven IJ\ ere actions of B cells
that develop millions of distinct antigenSUE::: ' ( antibody
molecules through the rearrangement of antibodyE1(OC( g genes and
the diversification of the sequences encoding : f 2 c'l-
ge'1-combining sites in the V regions of antibody molecules. - r6
'lumo-al response depends on activation by helper T (TH) cells, .',
([) oeoends, in turn, on their prior activation in the lymph nodes
art gen-presenting cell s, largely dendritic cells. The latter
process bl prole r s lOW oligopeptides t hat are recognized by the
T-cell receptors 01 Tli cells, whi ch proceed to activate B cells
that have developed, by chance, anti bodies that recognize the
oligopeptide antigens. T-cell receptors (TCRs) are used as well by
cytotoxic T cells (Tcl cells (also te rmed CTLs), which rely on
these receptors to recognize and kill target cell s displaying
cognate antigens. The activation of the Tc cells also depends on
prior stimulation by TH cells. A third class of T cells that also
expresses antigen-specific T-cell receptors are the regulatory T
cells, often called Tre9 s. These play important roles In
suppressing the actions of both Tc and TH cells and thereby 670
prevent inappropriate activation of immune responses that might
otherwise lead to a breakdown of tolerance and resulting autoimmune
disease. These various manifestations of adaptive immunity are
augmented by arms of the innate immune'system (right), specifically
cell types that can aid in the elimination of pathogens and cancer
cells without any prior" education" through previous exposure to
these entities. Thus, natural killer (NK) cells are primed to kill
many types of cancer cells because of the abnormal configuration of
cellsurface molecules displayed by these cells; macrophages are
also capable of recognizing and killing many cellular pathogens
without any prior exposure to these agents. Although macrophages
and NK cells cannot themselves recognize most cell-surface
antigens, the coating of potential target cells by antibody
molecules (produced by the adaptive immune response) will attract
macrophages and NK cells, which will use their Fe receptors to bind
to the constant (C) regions of antibody molecules and proceed to
kill the antibodycoated cells. Similarly, the complex group of
plasma proteins known as complement may also recognize antibody
molecules bound to the surface of a cell and then kill this cell by
inserting channels in its plasma membrane. Rejection of
histoincompatible tumors cells of various strains of mice display
distinct, genetically templated major histocompatibility (MHC)
molecules on their surfaces. (In this instance, however, it is not
the bound oligopeptide antigens that evoke an immune response but
the MHC molecules themselves, which vary slightly in structure from
one strain of mouse to another.) For example, engrafted cancer
cells from BALBI c mice were recognized as being of foreign origin
(and were therefore histoincompatible) when introduced into C57/BL6
mice, and vice versa (Figure 15.15). These graft rejections from
dissimilar, allogeneic (i.e., genetically distinct) mouse strains
were not observed when tumor cells of BALB/c origin were grafted
into BALB/c hosts, that is, into syngeneic hosts that, by
definition, shared an identical genetic background and identical
histocompatibility antigens with the engrafted cells. [In fact, the
term histocompatibility derives from the observation that tissue
("histo-") from mice of one inbred strain can be grafted and
established in the bodies of other members of the same genetic
strain and are in this sense "compatible."] The observed rejections
of allogeneic tumors represented a detour for the young field of
tumor immunology, since they shed no light on how the immune system
of a mouse or human host would recognize cancer cells that arise in
its own tissues. Still, this early work did make one profoundly
important point: in addition to eliminating microbes and various
types of viruses, the immune system is capable of destroying
mammalian cells that it recognizes as foreign or, quite pOSSibly,
as othervvise abnormal. As an additional corollary, these
observations of immune function led to the conclusion that cancer
cells could never be transmitted from one individual to another
(but see Sidebar 37 . , and Figure 15.16). An alternative strategy
was then embraced for studying the immunosurveillance problem. If
the immune system were indeed responsible for suppressing the
appearance of tumors, animals with compromised immune systems
should suffer increased rates of cancer. Such cancers, which
originated within their own bodies-so-called autochthonous
tumors-were, of course, of the same histocompatibility type as the
remaining tissues in these animals. In these situations, the issue
of histocompatibility (and -incompatibility) was rendered
irrelevant. 1( . carcinogen.. BALB!c ! tumors C57/BL6 ! tumors
tumo, celiAQ- tumm cell n inject cells inject cells Figure 15.15
Syngeneic mice and MHC variability The use of inbred strains of
mice has revealed that major determinants of the immunogenicity of
the cells of these mice (and of mammals in general) are the MHC
class I molecules. These molecules are highly polymorphic w ithin a
species, indicating that one individual (or one inbred strain of
mice) almost al ways has a different set of MHC class I molecules
from another (red, blue cell surface molecules) Therefore, if a
tumor arises within a BALB/C mouse, it is often transplantable into
a syngeneic host, i.e., another BALB/c mouse, but not into an rl rl
allogeneic host, such as a C57/BL6 tumors in no tumors in no tumors
in tumors in mouse. The converse is true for tumors syngeneic
allogeneic allogeneic syngeneic host host host host ariSing in
C57/BL6 mice. 671 --Chapter 15: Crowd Control: Tumor Immunology and
Immunotherapy Figure 15.16 Regression of CTVS and re-expression of
MHC antigens One exception to the rule of the nontransmissibility
of cancer from one organism to another comes from canine
transmissible venereal sarcoma; its cells are transferred from one
dog to another via sexual intercourse. The transferred cells
initially form a vigorously grovving tumor, which is, however,
rejected after several months (see also Sidebar 37 . ) (A) While
the canine transmissible venereal sarcoma (CTVS) cells initially
express very 10 levels of MHC class I proteins, aft er 12 eeks of
tu mor growth these proteins begin to be fe-expressed, as seen here
in t he tumors borne b hree dogs. (8) Thi s re-expression results,
at least in part, from s'gnals released by tumorinfiltra: 'og
lymphocytes (TILs) Fresh ( ulwre med'um has li ttle effect on the
express o ~ 0 HC class I (green) or class I red) prot eins by CTVS
cells (left) aaors re eased by TILs isolated from o'oc;:'esslog
tumors (1 st 12 weeks) also a e Ie effect (middle). However, "2SorS
released by TILs from regressing : VTlors (12-21 weeks) potently
induce He protein expression (right) by cultured CTVS cells. Such
re-expression appears to be responsible for tumor regression in
vivo. (From yw. Hsiao, K. W. Liao, S.w. Hung, and R.M. Chu, }
Immuno!. 172:1508-1514,2004) (A) (8) "D 25 40 Q) Vl -- ... Vl Q) 20
~ ~ 0.. -- 30 x ~ ~ Q) 15 V ' I ~ u ~ Ol 20I '+- . ~ ~ 10 o ~ '+0
;oR ~ 10o 0.. x 5 xQ) "D c: Q) 0.1 I + medium from TILs from
regressing tumors co-cultured with CTVS + medium from TILs from
progressing tumors co-cultured with CTVS dog 1 dog 2 dog 3 0 3 6 9
12 15 18 21 weeks after implantation + fresh medium In the late
1960s, immunocompromised mice of the Nude strain first became
available to cancer researchers. These mice lack a functional
thymus-the tissue in which the T lymphocytes of the immune system
initially develop. Their lack of hair, another distinct phenotype
of this strain, gave them their name (see Figure 3.13). The
research that followed in the early and mid -1970s revealed that
these mice are no more susceptible to spontaneously arising or
chemically induced autochthonous tumors than are their normal,
wild-type littermates. So, the immunosurveillance theory suffered a
major setback, having failed a major critical test of its validity.
It lost credibility and retreated from the main arena of cancer
research for two decades. But this rejection was premature. Only
years later did it become apparent that mice of the Nude strain,
while lacking many of their normal T lymphocytes, retain other
components of their immune system in an intact form. For example,
some types ofT cells may be able to develop outside the thymus, the
normal site of maturation of these cells. In addition, a very
important type of immune cell-the natural killer (NK) cell-is able
to develop totally outside the thymus, and thus NK cells are
present in large numbers in Nude mice. In the 1980s, researchers
began to accumulate evidence that NK cells are actually very
important in recognizing and killing a variety of abnormal cells,
including cancer cells. So in the end, the lessons taught by the
low cancer rates of Nude mice were of limited value, since these
mice did indeed continue to harbor functionally important
components of the inunune system. Still, Nude mice, as well as
other types of immunocompromised mice, have proven to be of great
value in cancer research (Sidebar 38 0 ). Evidence also began to
accumulate that certain chemically induced tumors in mice were
antigenic and could be recognized and eliminated by the immune
system. For example, in one set of experiments, cells from a
3-methylcholanthrene (3MC)-induced tumor were irradiated prior to
injection into mouse hosts in order to prevent the proliferation of
these cells in the hosts (Figure 15.17). (The chemically induced
tumor had been induced in the same strain of mice as these hosts.)
Subsequently, the mice received a second injection of live tumor
cells originating from the same tumor or from a second 3MC-induced
tumor; the cells originating from the same tumor did not grow,
while the cells from the second tumor did grow and form a new
tumor. This indicated that the two tumors were antigenically
different and that the initial exposure to dead cancer cells had
immunized the mice against live cells originating in the same
tumor. Hence, tumor cells could have distinctive antigens, and
under certain conditions, these antigens could provoke the immune
system to attack and kill such cells. 672 Cancer susceptibility and
immune function 15.8 Use of genetically altered mice leads to a
resurrection of the immunosurveillance theory In the mid-1990s,
several lines of research gave new life to the long-discredited
immunosurveillance theory. These experiments derived from the then
recently gained ability to create genetically altered strains of
mice at will. This technology (see Sidebar 7.10) was exploited to
create mice whose genomes lacked one or more of the genes known to
play critical roles in the functioning of the immune system. One
group of experiments used mice that were rendered incapable of
expressing the receptor for interferon-y (IFN-y) through targeted
inactivation of the responsible gene in their germ line. Like
growth factors, IFN-yis a diffusible protein factor that conveys
signals from one cell to another and induces responses in cells by
binding and activating its cognate cell surface receptor.
Importantly, IFN-y has not been found to be released by cells other
than those of the immune system. Consequently, any changes observed
following deletion of the IFN-y receptor gene from the mouse genome
could be attributed to defects associated with immune cells and
their interactions with the remaining cells in the body.
Strikingly, mice tha t lack the IFN -yreceptor in all of their
cells were found to be 10 to 20 times more susceptible to tumor
induction by the chemical carcinogen 3 -methylcholan th rene. In
another set of experiments, tumor cells were forced to express a
dominantnegative IFN-y receptor, rendering them unresponsive to the
IFN-y released by various types of immunocytes. These cells were
then injected into wild-type mice and found to be more tumorigenic
than tumor cells carrying the corresponding wild -type receptor.
This particular experiment suggested that the IFNY receptor
displayed by cancer cells enables them to respond to IFN -y
released by immunocytes, and that this response usually prevents or
retards the growth of tumors formed bv these cells. These striking
effects of IFN-y could be associated, at least in part, with the
actions of the natural killer cells. The NK cells were discovered
and named because of their innate ability to recognize tumor cells
as abnormal and to eliminate them. Once NK cells identify cancer
cells as targets for elimination, they release IFN -y in the
vicinity of the targeted cells. The released IFN -y, in turn,
elicits several distinct responses. As mentioned earlier, IFN -y
enables the NK cells to call in other types of immune cells to
assist in killing targeted cancer cells, thereby amplifying the
immune system's response. Among the responding immune cells are
macrophages, which aid not only by killing the cancer cells
directly but also indirectly, by functioning as professional
antigen-presenting cells (APCs) that process and display antigenic
molecules derived from the corpses of their victims (see Figure
15.6). At the same time, IFN-y stimulates targeted cancer cells to
display on their surfaces increased levels of class I MHC molecules
that may carry oligopeptide antigens capable of provoking further,
highly specific adaptive immune responses. This helps to explain
why transformed cells lacking the IFN-yreceptor are more
tumorigenic than counterpart cells that do display this receptor.
All of these responses seemed to be defective in genetically
altered mice lacking the IFN-y receptor; such mice were also found
to have an increased susceptibility to certain types of
spontaneously arising tumors. When taken together, these
experiments provided compelling validation of the idea that immune
surveillance plays a critical role in tumorigenesis, at least in
chemically induced tumors of mice. Further support of the
immunosurveillance theory came from mice that had been deprived of
the gene encoding perforin, the protein used by lymphocytes
irradiated tumor cells immunize mouse with irradiated tumor cells n
.. inject viable cells inject viable cells of the same tumor from a
second, independently induced tumorI ! ~ O host response host
response rejects tumor permits proliferation cells and of cells and
prevents tumor growth tumor formation Figure 15.17 Immunization of
mice by exposure to killed cancer cells Mice were initially
injected with irradiated, killed cancer cells (red) deriving from
one chemically induced tumor. When these mice were subsequently
injected with li ve cells from the same tumor, the cells failed to
grow (lower left). However, when these mice were injected with live
cells from a second tumor (blue), the cells proliferated and formed
a tumor mass (lower right). The reciprocal experiment (not shown)
yields the opposite results, i.e., injection with killed blue
cancer cells rendered mice immune to the blue tumor but not to the
red tumor. 673 Chapter 15: Crowd Control: Tumor Immunology and
Immunotherapy (A) (8) ta-chain locus T cell ~ - c h a i n locus
'1;, Figure 15.18 RAG proteins and TCR gene rearrangement The RAG-1
and RAG-2 proteins are responsi bl e for the rearrangement of DNA
segments that leads to the formation of both antibody molecules and
T-cell receptors (TCRs). (A) This diagram illustrates the
organization of the unrearranged genes encoding the a and ~ chains
of the TCR. The RAG-1 and RAG-2 proteins rearrange the germ-line
versions of these two genes through the deletion of intragenic
segments and the attendant fusion of previ ously di stantly linked
DNA segments within the a and within the ~ gene. Rearrangement
within the a gene is achieved when the RAG proteins juxtapose an L
segment encoding a leader sequence (L) at the N-terminus of the a
chain with one of the 70 to 80 Va gene segments (red) and one of
the 61 Ja segments (yellow); since the choice of individual Va and
Ja segments to be fused is essentially random, this results in a
large number of combinations of joined Va- Ja segments and a
comparably large number of distinct antigen-binding pockets. A
similar set of rearrangements occurring independently on a
different chromosome results in the formation of the ~
chain-encoding segments of the TCR. Since the antigenrecognition
domains of the TCR are created cooperatively by the amino acid
sequences encoded by the a and ~ chains, these RAG-1!2-mediated
rearrangements of the TCR-encoding genes are able to create a vast
number of distinct antigenbinding domains. (8) The TCR (above) on
the surface of a cytotoxic T cell (Te, CTL) enables the T cell to
recognize a specific oligopeptide antigen (yellow) displayed by an
MHC class I molecule (below; see also Figure 15.78) on the surface
of a potential target cell; such recognition and binding by the TCR
is achieved by its Va and Vp domains (colored loops) whose
generation is described in panel A. (Confusingly, the Va domain of
the TCR is created by juxtaposition of Va and J a DNA segments of
the a-chain locus, while the Vp domain is created by fusion of the
Vp, Dp, and Jp DNA segments of the ~ chain locus, all illustrated
in panel A.) Once such recognition has occurred, it may result in
the killing of the target cell (see Figure 1512). Since TCRs are
also used by T Hand Tregs for other immune functions (not shown;
see Figure 15.14), loss of TCRs caused by inactivation of a RAG
gene leads to crippling of many components of the multi-faceted
cellular immune response as well as the inactivation of the humoral
response, which dependends on RAG 1!2-mediated rearrangement of
antibody genes. (From c.A. Janeway Jr. et ai, Immunobiology, 6th
ed. New York: Garland Science, 2005.) and NK ceUs to mediate
killing of targeted cells. Recall that perforin is used by
cytotoxic cells to create channels in the plasma membrane of their
victims, allowing the entrance of apoptosis-inducing granzymes (see
Figure 15.12A). Mutant mice lacking the ability to make perforin
showed an elevated incidence of spontaneous tumors and were also
more susceptible to developing tumors following exposure to
3-methylcholanthrene. Similarly, increased cancer susceptibility
was registered in genetically altered mice that lacked the RAG-lor
RAG-2 proteins; these two proteins are responsible for rearranging
the genes encoding soluble antibody molecules as well as those
encoding the antigen-recognizing T-cell receptors (TCRs) displayed
on the surfaces ofT cells (Figure 15.18). Such RAG-lor -2-negative
mice lack T lymphocytes, B lymphocytes, y8 T cells (not discussed
further in this chapter), and a subclass of NK cells called NKT
cells. As a consequence, these mice have severely compromised
adaptive immune responses. ~ target cell 674 Cancer susceptibility
and immune function For example, in one experiment, 3-MC treatment
caused 30 of 52 RAG-2-i- mice to develop sarcomas, while only 11 of
57 wild-type mice of the same genetic background and treated in
parallel formed these tumors. The mutant mice were also found to be
far more susceptible to spontaneously arising cancers. Thus, 50% of
older (18-month-old) RAG-2-negative mice developed spontaneous
gastrointestinal malignancies-a twnor that is otherwise rare in
wild-type mice ofthis age. Arguably the most persuasive evidence
supporting the role of immunosurveillance in cancer prevention
comes from detailed studies of the 3-MC-induced sarcomas growing in
either R A G ~ i - or wild-type mice. When tumor cells prepared
from these two groups of sarcomas were grafted into new R A G ~ i -
hosts, both groups of sarcomas seeded tumors in these new hosts
with high efficiency (Figure 15.19). A very different outcome was
observed, however, when tumor cells were transplanted into
syngeneic vvild-type (and thus immunocompetent) hosts. Cells from
17 tumors that had previously been induced in wild-type mice all
succeeded in generating tumors in their new hosts. In contrast,
cells from 8 of 20 tumors that had previously been induced by 3-MC
in RAG2-i- mice failed to form tumors, being rejected by the immune
systems of these wild-type hosts (see Figure 15.19). These
observations open our eyes to an entirely new dimension of tumor
immunology. They suggest that when 3-MC-transformed cells arise in
an immunocompetent host, those that happen to be strongly
immunogenic (and thus capable of provoking some type of immune
response) are effectively eliminated by the host, resulting in the
survival and outgrowth of only those cancer cells that happen to be
weakly imm,unogenic. The latter then multiply and form tumors in
their original hosts and, later on, succeed in doing so when
transplanted into other immunocompetent hosts. Hence, these tumors
represent a subset of those that originally arose in the primary
hosts. The missing, strongly immunogenic twnors are apparently
eliminated early in tumor progression by host immune systems and
therefore never see the light of day (see Figure 15.19). In
contrast, when 3-MC-transformed cells arise in an immunocompromised
host (see Figure 15.19A), two classes of tumors are initially
formed, as beforethose that are strongly immunogenic and those that
are weakly immunogenic; both types of tumor cells survive in an
immunodeficient host. Later, when these tumors are transplanted
into immunocompetent hosts, those that are strongly immunogenic
fail to form tumors, while those that are weakly immunogenic
succeed in doing so. We conclude that in wild-type mice, a
functional immune system plays an important and effective role in
eliminating a significant fraction of the tumors that are initially
induced by 3-MC. These observations indicate that the immune system
of these mice plays an active role in determining the identities of
tumors that arise and the antigens that they express. This active
intervention in the phenotype of tumors has been termed
immunoediting, to indicate the weeding out of some tumors and the
tolerance of others. Immunoediting can be thought of as a type of
Darwinian selection, in which the selective force is created by the
directed attacks of the immune system on incipient tumors. 15.9 The
human immune system plays a critical role in warding off various
types of human cancer Because the biology of mice and humans
differs in so many respects, we need to interpret the results
described above with caution when attempting to understand the role
of the human immune system in defending us against cancer. In 675
Chapter 15: Crowd Control: Tumor Immunology and Immunotherapy
addition, the chemical carcinogens used in the experiments
described above may well create tumors in mice that are far more
antigenic or immunogenic than spontaneously arising human tumors
(to be discussed in Section 15.12). Figure 15.19 Effects of immune
function on the development of anti-tumor immune responses Both
wild-type (wt) and RAG2-1immunocompromised mice were exposed to the
potent carcinogen 3-methylcholanthrene (3-MC) (A) When the tumors
induced in the RAG2-/- mice were transplanted back into RAGL'-
hosts, they all formed tumors (above). However, when the tumors
induced in the RAG;z--I- mice were transplanted back into wild-type
hosts, 8 of 20 tumors failed to form (below) Each line presents the
kinetics of growth of a single implanted tumor. (B) This experiment
and other experiments using tumors induced in wt mice (not shown)
are summarized here. Following exposure to 3-MC, the wt mice
developed fewer tumors (blue) than did the RAG2-/- mutants (blue
and red) The tumors from the two groups of mice were excised and
cells from each were converted to a cell line that could be
propagated in vitro. Cells from each of these cell lines were then
transplanted back into either wt mice or RAG2-/- mutant mice. Cells
from all of the tumors that appeared initially in the wt mice (bl
ue) were able to form new tumors in both w t and RAGL'- hosts
(left). However, cells from all of the tumors that arose and grew
initiall y in the RAGL'mice were able to form new tumors in RAGL'-
mice (red, blue), but only some of these (blue) were able to form
new tumors in the wt mice (right). These experiments suggested that
3-MC initially induced t wo types of tumor cells in all of the mice
strongly immunogenic (red) and weakly immunogenic (blue). Both red
and blue cells formed tumors in the RAG2-/- mice, but only blue
cells formed tumors in the wt mice, since any initially formed red
tumor cells were eliminated by the functional immune systems of
these mice. This meant that the tumors that did arise in wt mice
were already selected for being weakly immunogenic and thus capable
of forming new tumors in other wt mice. (A, from V. Shankaran, H.
Ikeda, AT Bruce et al., Nature 410 1107- 1111, 2001) (B) (A) 25 20
15 E.s 10 ~ o ~ , . E t :, .. ". 3 20" : :;, " .. / c.. mice
engrafted with 24D3 lymphocytes"" ::J::J::J a:l tumor cell Iysates
figure 15.23 Specificity of the antigen display by a chemically
induced fibrosarcoma Mice of the BALBIc strain were immunized with
Iysates of cells of the 3-methylcholanthrene-induced Meth A
fibrosarcoma line. A line of antigen-presenting lymphocytes, termed
24D3, was developed from these mice. (A) As gauged by their
incorporation of 3H-thymidine, proliferation of these lymphocytes
in culture was stimulated by addition of Meth A cell Iysates (2nd
bar). However, Iysates prepared from 14 other tumor cell lines,
including other methylcholanthrene-induced sarcomas, UV-induced
squamous cell skin carcinomas, lymphomas, a melanoma, and a lung
carcinoma, failed to stimulate proliferation of these lymphocytes
(remaining channels). In the absence of antigen (1st channel) no
proliferation was seen. (B) When a clonal population of these 24D3
lymphocytes was introduced into BALBIc hosts, the formation of
tumors by subsequent ly injected Meth A fibrosarcoma cells was
fully blocked (left panel). However, the formation of tumors by an
unrelated fibrosarcoma line, termed CMS5, was unaffected by the
presence of these lymphocytes (right panel). (From T. Matsutake and
P.K. Srivastava, Proc Natl. Acad. Sci. USA 983992-3997, 2001) 686
Tumor-associated transplantation antigens their descendants
expressing low levels of TSTA will survive long enough to be
studied by an experimenter, greatly complicating the biochemical
isolation and identification of the TSTA protein. In recent years,
several of the genes encoding 3-MC-induced TSTAs have nevertheless
been isolated by gene cloning procedures. In one cloning strategy,
the oligopeptides that were bound to MHC class I molecules on the
surfaces of 3MC-transformed cells and served as targets for immune
recognition were eluted from the MHC molecules, purified, and
subjected to amino acid sequencing. The resulting amino acid
sequences were then used to predict the nucleotide sequences of the
encoding genes, which made possible the cloning of these genes.
Sequence analyses of the TSTA-encoding genes cloned from these
tumors showed that the genes were all point-mutated alleles of
normal cellular genes encoding various cellular proteins, none
involved in any obvious way in the transformation of these cells
(Sidebar 41 . ). . These observations suggest that during the
course of chemical carcinogenesis, the 3-MC carcinogen, a known
point mutagen (Section 12.6), mutates both a proto-oncogene (often
the K-ras gene) in target cells and additional genes that, as
mutant alleles, specify TSTAs; the latter genes are struck at
random-innocent bystanders tl1at play no causal role in
tumorigenesis but happen to have been damaged by the large doses of
mutagenic carcinogen used to provoke tumor formation. Importantly,
the behavior of these chemically induced TSTAs is quite different
from that of the ISlAs resulting from tumor virus infection. For
example, SV40 virus can be used to induce a sarcoma in a mouse.
Subsequent removal of this SV40-induced sarcoma will result in
immunization of the mouse against subsequently inoculated tumor
cells that derive from this particular SV40-induced sarcoma as well
as a!l y other tumors that have been induced by SV40 virus. In this
instance, there is indeed a cross-immunity established, in that all
the SV40induced tumor cells seem to share a common TSTA or set of
TSTAs. It happens that the dominant ISlA responsible for this
cross-immunity is a familiar protein: it is the virus-encoded large
T oncoprotein, which is expressed at significant levels in all SV40
virus-transformed cells. This contrasts with the behavior of a
group of 3-MC-induced cancers, where each tumor expresses its own
unique TSTA or set ofTSTAs. Observations like these raise the
question whether similar mechanisms operate during human
tumorigenesis. Thus, do the highly mutable genomes of cancer cells
(Chapter 12) generate mutant, antigenic proteins as inadvertent
by-products of the mutagenesis that drives tumor progression (see
Sidebar 15. I)? Or are the 3-MC-induced TSTAs artifacts of the high
doses of mutagenic carcinogens used in many mouse tumorigenesis
experiments that do not accurately reflect the mutagenic processes
that create human tumors? 15.13 Tumor-associated transplantation
antigens may also evoke anti-tumor immunity As noted above,
tumor-associated transplantation antigens (TATAs) represent normal
cellular proteins that, for one reason or another, have failed to
induce tolerance. When these normal proteins are expressed by
tumors they evoke a measurable immune response, often involving
both the humoral and cellular arms of the immune system. For a
variety of reasons, the antigenicity of melanomas has been more
intensively studied than that of all other human tumors (Sidebar 42
. ). Much of their antigenicity stems from their display of certain
TATAs. Melanoma cells may Sidebar 15.1 Microsatellite instability
often leads to more immunogenic tumors As described in Section
12.4, defects in the DNA mismatch repair machinery create the
condition of microsatellite instability (MiN), which leads to
mutations accumulating in hundreds, possibly thousands of cellular
genes within tumor cell genomes. Among other consequences, these
mutations generate shifts in the reading frames of many of these
genes. The resulting mutant alleles often encode novel amino acid
sequences, sometimes termed "frameshift pep tides," some of which
may function as potent tumor-specific antigens. This logic predicts
that the 15% of human colorectal cancers that exhibit
microsateJlite instability should interact with the host immune
system differently from the majority of colorectal cancers that
show no MIN and instead exhibit chromosomal instability (CIN). In
fact, the MIN tumors show a markedly higher degree of
tumor-infiltrating lymphocytes (TILs) and a lower degree of
metastasis. Moreover, antigen presentation by their MHC class I
proteins . is compromised Jar more frequently than in colorectal
tumors with chromosomal instability (60% vs. 30%), suggesting that
the MIN tumors are under greater pressure to evade killing by
various arms of the immune system, and that they undertake the
immunoevasive maneuver of blocking antigen presentation by their
cell-surface MHC class I proteins. Taken together, these
observations suggest that the MIN that enables some colorectal
tumors to evolve more rapidly can exact a price in the markedly
higher immunogenicity of these tumors. 687 Chapter 15: Crowd
Control: Tumor Immunology and Immunotherapy Figure 15.24 Normal
proteins displayed as tumor-associated antigens on melanoma cells
(A) The tyrosinase enzyme, which is involved in pigment production
in melanocytes and melanomas, is detected here in the melanocytes
of the skin (red, located Just above the basement membrane in the
skin) through the use of a monoclonal antibody; it is not
detectable in other normal tissues. Its expression by melanoma
cells can cause them to become immunogenic and the target of ki lli
ng by cytotoxic lymphocytes. (B) The spermatogonia in the testis
have been stai ned here with a monoclonal antibody agai nst the
MAGE-l antigen; normall y this antigen is seen onl y in the
placenta. Its expression has been detected in a variety of human
tumor types and has been studied in detail in melanomas because it
is often immunogenic when expressed by these tumors. (A, from VT.
Chen, E. Stockert, S. Tsang et al., Proc. Nat!. Acad. Sci. USA
928125-8129,1995; B, from JL Cheville and Pc. Roche, Mod. Pathol.
12974-978, 1999 ) overexpress certain proteins that are present in
their normal melanocyte precursors, albeit at lower levels. Such
lineage-specific proteins are sometimes cal.Ied differentiation
antigens, implying that their display is a vestige of the
differentiation program that previously governed the behavior of
the normal cellular precursors of tumor cells. Included among the
melanoma TATAs are transferrin, tyrosinase (Figure 15.24A), gpl00,
Melan-A/MART-l, and gp75. The display of these differentiation
antigens by melanoma cells often provokes a vigorous response by
the immune system, which results in a very peculiar form of
autoimmune disease-vitiligo-the depigmentation of large areas of
skin seen in some melanoma patients (Figure 15.25) . This
depigmentation is a specific response to the presence of a
melanoma. For example, when 104 renal carcinoma patients were
treated with the cytokine interleukin-2 OL-2) in order to enhance
their anti-tumor immune responses, none developed vitiligo; in
contrast, of 74 melanoma patients who were treated Similarly, 11
developed vitiligo. In these melanoma patients, it is clear that
the immune response provoked by the melanoma TATAs leads, as a
by-product, to attack and destruction of normal melanocytes, which
also express these antigens. This type of vitiligo is formally
analogous to the paraneoplastic syndromes discussed earlier
(Sidebar 400), in which the display by tumors of cellular proteins
results in the destruction of normal tissues that also happen to
express these proteins. Significantly, melanoma patients showing
vitiligo usually survive for longer periods than those who
don't-suggesting that their immune systems are effective in
controlling the melanomas, at least for a period of time. (For
example, in a large population of melanoma patients described in
1987, 75% were still alive five years after initial diagnosis;
among the subgroup of these patients who exhibited concomitant
vitiligo, 86% survived for this period of time.) The antigenicity
of human melanoma cells may also derive from their display of the
other major subclass of TATAs, the oncofetal antigens-literaHy
those antigens that are displayed during embryogenesis and once
again by tumors. Included among these are the antigens called
either cancer germ-line or cancertestis (eT) antigens, to reflect
their normal e>-''}Jression in the germ cells of the testis and
the fetal ovary. The genes for a number of these antigens, such as
(A) tyrosinase antigen (8) MAGE-1 antigen r--.............. skin
melanocytes testes spermatagonis 688 Tumor-associated
transplantation antigens Figure 15.25 Autoimmune depigmentation
provoked by melanomas The melanoma patient shown here, who was
dark-skinned prior to the onset of melanoma, has lost almost all of
his skin pigment except for several isolated areas (face, armpit)
due to the autoimmune attack inci ted by the melanoma cell s. The
condition of pigment loss, known as vitiligo, is often correlated
with a longer survival of melanoma patients. (Courtesy of A.N.
Houghton.) MAGE-1 (Figure 15.24B), MAGE-3, BAGE, GAGE-I, and
GAGE-2, have been cloned (Sidebar 43 0 ). By 2003,44 genes or gene
families encoding a total of 89 distinct cancer-testis antigens had
been identified. As an aside, it seems that the absence of immune
responses against these antigens in males is likely due to the fact
that several of the cell types in the testes do not express MHC
class I molecules and are thereby prevented from presenting their
internal contents to the immune system. (Of course, females may
never express these