A NOVEL ASSAY FOR SERUM FERRITIN
Meg Isabel Langfur
B.A., University of Michigan, 1983
A thesis submitted to the
University of Colorado at Denver
in partial fulfillment
of the requirements for the degree of
Master of Basic Science
1999 by Meg Isabel Langfur
All Rights Reserved.
This thesis for the Master of Basic Science
Meg Isabel Langfur
has been approved
Langfur, Meg I. (Master of Basic Science)
A Novel Assay for Serum Ferritin
Thesis directed by Professor Douglas Dyckes
Previous studies have confirmed the presence of a novel glycosylated ferritin,
different from cellular ferritin [Campbell et al., (1993) J Cell Biochem. 53 420-423].
From this we infer that this glycosylated ferritin has characteristics that can be
exploited in the development of a radioimmunoassay. Numerous RIA, or other
immunoradiometric assays based on monoclonal and polyclonal ferritin antibodies
are currently available on the market. Without exception, however, these products
are grown against cellular ferritin antigens. The consequence of this preparation, is a
significantly less pure antibody with much less specificity for the antibody.
We now report the production of a highly pure serum ferritin antigen, which
was utilized to grow a polyclonal a-serum ferritin antibody. Western blot
electrotransfer showed high specificity of this antibody to the antigen and minimal
cross reactivity to the cellular ferritin moieties. With this foundation in place, we
report the development of initial phase studies of an RIA using the purified antigen,
125I labeled ferritin, and the purified antibody. With further study, it is our hope that
this RIA will permit the clinician to distinguish between glycosylated and cellular
ferritin in serum. Previous studies have shown this to be a reliable correspondence to
This abstract accurately represents the content of the candidates thesis. I recommend
I dedicate this thesis to my mother, Rosalyn Langfur. Without her unfaltering
support, her understanding, her patience, and her humor, this would never have been
finished. I love you Mom.
My thanks to staff at the Graduate School at the University of Colorado at Denver for
their support and patience. My appreciation for my colleagues at the University of
Colorado Health Sciences CenterDivision of Hematology for their assistance and
their sense of humor, especially Valarie Allen, without whose tireless editing work, I
would have been lost. My admiration for my advisor, Dr. Doug Dyckes for his wise
guidance these past years. My gratitude to Rhoda Schleicher for her help in the
laboratory and her friendship. My love to Dad, Hal, Susan, Rio, and Simon; thank
you for watching over me.
1. Introduction and Background...........................................1
1.1 Human Blood: The Study Environment....................................1
1.2 Iron Transport in the Blood..........................................2
1.3 Control of Body Iron Stores...........................................3
1.4 Ferritin Synthesis....................................................4
1.5 Ferritin Form.........................................................4
1.6 Free Iron: Ferritins Protective Role................................6
1.7 Structure and Function of Ferritins in the Serum......................10
1.8 Ferritin in Disease States...........................................12
2. Problem Statement.....................................................20
2.1 The Clinical Use of the Ferritin Assay...............................20
2.2 The Study and Assay Development......................................22
3. Materials and Methods.................................................23
3.1 Purification of Human Serum Ferritin.................................23
3.2 Concanavalin-A Chromatography........................................24
3.3 Protein Assay........................................................25
3.4 Anti-Ferritin Immunoaffinity Chromatography.......................25
3.5 Preparation of Anti-Human Serum Ferritin Antiserum................26
3.6 Sephacryl Gel Exclusion Chromatography............................26
3.7 Polyacrylamide Gel Electrophoresis................................27
3.8 Western Blot Electrotransfer......................................28
3.9 Radiolabeling of SFt..............................................29
3.10 Double Antibody Radioimmunoassay.................................30
4.1 Results and Discussion.............................................31
4.2 Purification of SFt............................................. 31
4.3 Comparison with Commercial Antigens and Kits......................46
5. Summary/Future Studies..............................................51
1.1 -- Body iron Supply.................................................3
1.2 -- The 24 subunit molecule..........................................6
1.3 -- Cellular Injury from Iron Overload...............................9
1.4 -- SFt. Levels Vs. Hodgkins Disease...............................19
4.1 -- Elution Profile for Crude Purification..........................32
4.2 SDS-PAGE Crude Purification.....................................33
4.3 -- SDS-PAGE Complete Purification................................ 35
4.4 Elution Profile Concanavalin-A purification.....................37
4.5 Elution Profile for a-SFt purification..........................40
4.6 -- Western Blot of a-SFt and a-CFt................................42
4.7 - Diagram of Monoiodotyrosine................................... 44
4.8 Elution Profile of 1251 SFt purification........................46
4.9 SDS-PAGE of Commercial Kits.....................................48
4.10 SDS-PAGE of Commercial Kits...................................48
4.11 SDS-PAGE of Commercial Kits....................................48
4.12 Western Blots of Commercial Kits vs. a-SFt....................49
4.13 -- Western Blots of Commercial Kits vs. a-CFt....................49
1.1 -- Variation of SFt levels with Disease..............................15
1. Introduction and Background
1.1 Human Blood: The Study Environment
The total blood volume in a human adult is five to six liters, or 7 to 8 % of the
total body weight. Approximately 45% of the blood is composed of formed elements:
red blood cells (rbc), white blood cells (wbc), and platelets. The rbc or erythrocytes
contain hemoglobin which binds and carries oxygen; the wbc, or leukocytes, defend
the body against foreign non-self substances such as infection; and the platelets or
thrombocytes primarily function in the stoppage of bleeding. The remaining 55% are
the fluid portion of the blood. 90% of the fluid is composed of water and 10% is
proteins such as albumin, globulin, ferritin, plus carbohydrates, vitamins, hormones,
enzymes, lipids, minerals and salts, as well as waste material such as urea , Each
day almost 200 billion non-nucleated rbc are produced in the normal adult to replace
a like number reaching the end of their life span. Each red cell, contains more than a
billion atoms of iron, four held in each tetrameric molecule of hemoglobin, so that
more than 200 x 10 atoms of iron are needed daily for erythropoiesis, the formation
of rbc. This is almost 20 mg of iron by weight. Human iron metabolism is
distinguished by an efficient cycling of iron from recently destroyed to newly formed
rbc. Since less than 0.05% of the total body iron is acquired or lost each day, humans
are unique among animals in their effectiveness of iron conservation .
1.2 Iron Transport in the Blood
Iron is shuttled in the body via the transport protein transferrin. This single
polypeptide (MW approximately 80 kD), has two iron binding sites with which it is
able to move iron to cells requiring the metal primarily for hemoglobin synthesis.
Cells express transferrin binding sites when the uptake of iron is required and
suppress this expression when the need is satisfied. Once bound to the appropriate
cell-surface receptor, the transferrin-iron complex is moved into the cell through
receptor-mediated endocytosis. Inside, the iron is released and the transferrin is
recycled to the surface where it disengages [1,2]. Regulation of the transferrin is
dependent upon cytosolic concentrations of iron and activated by an iron responsive
element binding protein (IRE-BP) situated on the 5 untranslated end of the
transferrin mRNA. In low iron conditions, the IRE-BP disassociates from the 5 end
and attaches to the 3 end of the transferrin receptor mRNA. This effectively
lengthens the half-life of the receptor mRNA, which results in the synthesis of
additional receptors that can accept transferrin-iron complexes [3,4].
1.3 Control of Iron Body Stores
The disproportionate distribution of iron stores in the human body is
represented by the following schematic:
Fig. 1.1 Body iron supply. Schematic representation of the
movement of iron in the adult human. Circle area is proportional
to the iron concentration. Tf is the abbreviation for transferrin.
Adapted from Iron Metabolism in Health and Disease bv JS Brock
et al., WB. Sanders Co.,1994.
The concentration of iron in the human body is normally about 40 mg Fe/kg
body weight in women and 50 mg Fe/kg body weight in men .
Body iron stores are tightly controlled in the healthy individual and ferritin
plays an important part in this regulation. When an influx of free iron is required for
hemoglobin or metalloenzyme synthesis, intestinal mucosal cells and macrophages of
the reticuloendothelial system have as their primary function the release of iron
harvested from digestive tract contents and senescent red cells. When the uptake is in
excess of the metabolic requirements, cellular ferritin provides a storehouse for the
metal which can be accessed in times of increased iron needs .
1.4 Ferritin Synthesis
The liver, which is the bodys major iron storage organ, synthesizes ferritin
for the sole purpose of iron storage . The synthesis occurs on free polyribosomes
of the hepatocytes and on the rough endoplasmic reticulum . Iron is taken up by
the ferritin protein in the ferrous (Fell) form but the metal is oxidized at sites in the
channels (described below) to the ferric (Felll) form before it reaches the core. The
nature of this oxidation is still unclear, but spectroscopic and crystallographic studies
point to the involvement of the carboxylic residues which line the channels, as the
probable site of oxidation .
1.5 Ferritin Form
There are three documented types of ferritin: cellular or tissue ferritin (mw
450,000 Da), glycosylated heart ferritin (mw 140,000 Da), and blood serum or serum
ferritin (mw 139,000 Da). Ferritin is a unique protein in that it is unusually stable due
in part to the presence of disulfide bonding. It has high levels of hydrophobicity that
promote permanence in the tertiary and quaternary conformations . These qualities
allow ferritin to retain its biological conformation under stresses, such as high
temperature or acidity that quickly and efficiently denature less sturdy proteins .
Most mammalian ferritin subunits are designated either H (heavy) or L (light).
The tissue source determines the ratio of L to H subunits. Each of these subunits is
specified by its own gene, and each gene is situated on a different chromosome. The
H subunit gene resides on chromosome 11 and the L subunit gene resides on
chromosome 19 . The transcriptional complexity of this arrangement is obvious.
A cDNA for a third type of subunit, with intermediate mobility on SDS-Poly
Acrylamide Gels, is designated M and has been identified in the immature
erythrocytes of the bullfrog tadpole . Because the ferritin subunits from the
bullfrog and the parasitic worm Schistosoma mansoni resemble H-chains more than
they do L-chains , it has been suggested that the H-chain is the evolutionary
forerunner of the ferritin subunits thus far identified . Linder and Moroz [12,13]
have recently reported some atypical ferritins and ferritin subunits.
The predominant type, tissue ferritin, has been extensively studied and is
composed of a shell of 24 subunits surrounding a core of hydrous ferric oxide (FeO
(OH)) that can contain multiple atoms of free iron as shown below. Distinctive
channels have been shown through x-ray analysis to exist along a four-fold axis of the
molecule providing access to the central cavity for iron and other molecules <1.3 nm
in diameter .
Figure 1.2 Adapted from Ford et al., (1984). Schematic shows a cross section of
protein shell made up of 24 subunits. There are 6 iron channels allowing uptake and
release to the cytosol. Tissue ferritin has a mass of 450,000 daltons and the core may
hold up to 4,500 Fe atoms.
1.6 Free Iron: Ferritins Protective Role
Most organisms, including humans, require environmental iron for growth and
adequate metabolism. Small variations in the cytosolic concentrations of iron can
result in cell death and tissue damage. Researchers have connected high iron levels to
atherosclerosis, cancer, heart disease, arthritis, and accelerated aging , Ferritin is
the cellular protein responsible for the careful regulation of free iron. It functions as
the site for iron storage and detoxification . Without this particular function, free
cytosolic iron, some of which is present due to the destruction of phagocytes, can
initiate the autoxidation of ferrous iron in the presence of hydrogen peroxide. Called
the Fenton reaction, this process results in the production of the reactive hydroxide
radical and creates a dangerous scenario. Transition metals, such as the circulating
free iron act catalytically in free radical reactions. The Fenton reaction is the main
route in biological systems for producing the hydroxyl radical. The formation of this
radical is thermodynamically favored over the production of the primary reduction
intermediate, the superoxide anion radical. Ferrous iron along with hydrogen
peroxide and the hydroxyl radical can create DNA nicking on plasmids, which are
small circular DNA molecules that float free in the cytoplasm . At that point
inaccuracies in transcription and translation may result.
Clinical evidence for tissue damage caused by iron excess has been
documented in numerous studies of patients with hemochromatosis or iron overload.
Bomford and Williams (1976), Powell (1980), Niederau (1985), and Bassett (1986)
are just a few who have examined the correlation between hepatic cell iron
concentration and liver injury . Therapeutic reduction of iron levels by either
phlebotomy or the use of iron chelators has been shown to improve the status of the
hemochromatosis patient. Bradbear (1985), et al, reported in an article published in
the Journal of the National Cancer Institute in 1985 that clinical evidence proves an
association between hepatic fibrosis and cirrhosis due to chronic iron overload in
hemochromatosis and development of hepatocellular carcinoma . The
mechanism that charts the progression from cellular injury to neoplasm is not yet
clearly understood, but the risks to the patient of excess iron are undeniable.
Excess iron is implicated in so-called iron-induced peroxidative injury to the
polyunsaturated fatty acids that compose the phospholipid membranes of the cells.
In vitro studies have illustrated the volatility of ionic iron and its role in lipid
peroxidation. Subcellular organelles that are incubated with ionic iron show
extensive structural and functional damage as the peroxidation decomposes the
membranes. Lysosomes, microsomes, and mitochondria, as well as the endoplasmic
reticulum and nucleus, all show the consequences of iron-induced lipid peroxidation
in alterations of membrane permeability, lessening of the activity of membrane-
dependent enzymes, and the toxicity of lipid peroxides and their breakdown products.
The schematic shown below, gives a representation of the injury that might occur if
toxic levels of iron were to build in the tissue .
Figure. 1.3 Adapted from R.S. Britton, et al. Gastroenterology (T9911. The
schematic shown above proposes a pathway of cellular injury in iron overload.
While iron has been proven to initiate lipid peroxidation, the origin of this
initiation is not known. Histopathologic studies have identified excess cellular
ferritin in the liver tissue of hemochromatosis patients, and indications are that the
cellular ferritin CFt may release a small amount of iron in a form that may commence
lipid peroxidation. Generally, the presence of CFt is considered to be cytoprotective
in that the CFt stores most of its iron in a form that makes it unable to act as an
initiator. It has been suggested, however, that in times of severe and chronic iron
overload, this protective mechanism of the cellular ferritin may become overwhelmed
due to an increase in the intracellular pool of iron. Components of this pool may be
catalytically active in the stimulation of the lipid peroxidation . In the aerobic
environment of the human body, the protein ferritin is an ideal example of the
connection between structure and function, and the protein is essential in all iron
dependent organisms .
Very little is known about in vivo release of the iron stores contained in the
ferritin core. In vitro studies have shown that the release requires reducing agents,
free radicals or low pH . Iron release in vivo may occur in the lysosomes or
endosomes that have acidic pH. It is certain that the iron must be reduced in order to
be released from the ferritin molecule. The presence of superoxide radicals generated
by FMNH2 may catalyze the reduction of the ferric iron. Flavins such as riboflavin
may also play a role in the conversion of Fe (III) to Fe (II). Studies have shown that
in vitro iron release from ferritin in the presence of riboflavin reaches a maximum
rate of 100 atoms per minute per ferritin molecule. This means that a typical ferritin
molecule containing 4500 atoms of iron could release its entire store in under 60
minutes in response to an acute iron deficiency (anemia) .
1.7 Structure and Function of Ferritins in the Serum
Smaller amounts of ferritin are found in the blood serum but, beyond its
approximated molecular weight and the presence of some higher weight glycosylated
subunits, very little is known about the primary structure, site of synthesis, regulation
or metabolic function of serum ferritin [18,20,21]. Atomic absorption studies of the
serum protein from patients exhibiting the symptoms of iron overload show the serum
variety has a much lower iron concentration than that of cellular ferritin. The
isoferritins (serum ferritin found in the blood and CFt) show 0.02- 0.07fig Fe/pg
protein and a 0.2pg Fe/p g protein respectively . In this study, and later
experiments, we found there to be a complex mixture of ferritins in the serum. This
mixture purifies according to conventional protocols (ie., concanavalin-A
chromatography), shows similar molecular weights, different banding patterns on
SDS-PAGE, and cross-reacts to a-SFt (anti-serum ferritin antibody), produced in this
study (data not shown). For this reason, we will refer to the mixture of ferritins found
in the blood as SFts and refer to the serum ferritin purified in this study as SFt.
The presence of glycosylation on the serum ferritin molecule suggests that the
protein is secreted, possibly from phagocytes after degradation of hemoglobin .
While erythrophagocytosis may explain some of the serum ferritin concentrations, the
direct release of CFt through damaged cell membranes should also be considered. It
should be noted, however, that ferritins released from damaged parenchymal cells are
not glycosylated and will not bind to the lectin Concanavalin-A, commonly used to
purify glycoproteins, and could not, therefore be independently analyzed.
1.8 Ferritin in Disease States
Serum ferritin concentrations normally lie in the 15-300 jag /L range and
typically run lower for children and women. These concentrations are of tremendous
value in determining overall iron stores. Serum ferritin levels decline with Fe storage
depletion: a serum ferritin level below 12 pg /L is virtually diagnostic for absent iron
stores . Numerous disorders have also been shown to elevate the serum ferritin
levels independently of the body iron stores. Hemochromatosis, hepatitis B infection,
Hodgkins Lymphoma, Leukemia, breast cancer, rheumatoid arthritis, and
neuroblastoma [22,23,24,25,26] are just a few of the illnesses that show-serum ferritin
to be an acute phase reactant. SFts syntheses are increased in these pathologies as part
of a nonspecific response to the normal systemic effects of inflammation .
Defective clearing of circulating ferritins may also raise serum ferritin levels.
Malignancies are implicated in the elevation of serum ferritin levels as well. This may
be due to the release of CFt from damaged cell membranes, which is often seen
during the advance of neoplastic disease. In those patients undergoing chemotherapy,
serum ferritin levels are high in connection with cell lysis by the chemotherapeutic
agents . Hepatic dysfunction may impede clearing of the ferritin, and direct
synthesis by tumor and other cells may also account for elevated serum ferritin levels.
Because so little is known about SFts, however, it is difficult to make assumptions
about the mechanisms by which its concentrations increase during disease .
Beyond the diagnostic significance of correlating SFts levels with total body
iron stores, does the measurement of the protein offer any connection with disease
stage or progression? As mentioned previously, the SFts variety shows a
preponderance of the H subunit and G (glycosylated) subunits. Ion exchange
chromatography shows this form to be acidic with lower isoelectric points than the
cellular form, which tends towards basicity and higher isoelectricity, due to a high
proportion of L subunits. Isoelectric focusing based on the ratio of H to L subunits
with the relative ratio of G subunits to H subunits considered concurrently, allows for
detection of microheterogeneity. This detection allows for the diagnosis of
hemochromatosis versus [J-thalassemia (secondary iron overload) versus cytolytic
hepatitis. The acid/base ratio is decreased in idiopathic hemochromatosis, but rises in
thalassemia patients. The cytolytic hepatitis patient shows a different profile with
basic ferritin subunits in much greater concentration . The usefulness of
separating and quantifying serum ferritin levels begins to show its value.
The elegant mechanisms that the human body employs to fight disease have as
their foundation the leukocytes or white blood cells mentioned earlier. Three types of
white blood cells circulate to provide the first line of defense against non-self
invaders. Granulocytes function mainly to phagocytize bacteria. Nucleated
leukocytes migrate into the connective tissue where they build antibodies to both
bacteria and viruses. Monocytes attack non-self organisms not destroyed by the
granulocytes or leukocytes.
When the immunological defenses of the body are overwhelmed, the blood
continues to serve a function as a messenger from which diagnosticians can glean
valuable information concerning the progress of the illness. The diagnosis of anemia
and cancers, for example, are two of the familiar ways that the blood is used as a
reflection of disease. Of specific interest to this researcher is the measurement of the
iron storage protein ferritin found in blood serum and its correlative importance as a
disease marker. It is known that this form of ferritin, referred to previously as SFts, is
secreted as an acute-phase reactant . In the absence of inflammatory disease,
SFts levels can be related to tissue iron stores and be used to assess their size . In
chronic inflammatory disease, however, this relationship is lost, and SFts levels rise
Anemia, which arises from several possible etiologies, can be either acute or
chronic and may lead to a deficiency in stored iron that subsequently inhibits the
synthesis of hemoglobin. The characteristic iron deficiency of anemia is typically
shadowed as well by a decrease in SFts levels (to 12 ng/mL or less) . On the
other end of the spectrum is the autosomal recessive disorder, hemochromatosis.
Blood iron levels are tightly controlled and precisely balance dietary absorption with
loss due to sloughing of intact cells containing iron. In hemochromatosis, the
regulation of the absorptive step is disrupted by an HLA-linked locus on chromosome
6. This disorder of iron metabolism is identified by clinical symptoms such as
hepatomegaly, skin hyperpigmentation, cirrhosis, arthropathy, and diabetes mellitus.
Quantitatively, the disorder is confirmed by the measurement of SFts levels as a
relative index of body iron stores . Analysis of SFts levels also permit detection of
presymptomatic individuals, as well as the magnitude of transferrin saturation and
total iron binding capacity, which aid in assessing disease stage.
The table summarizes the variation of SFts levels some diseases:
Group Median Value (ng/mL) SFts
Adult Males 69-140
Adult Females 34-39
Iron Deficiency 1.5-5
Hemochromato sis 1131-3936
Adult Inflammation 213-250
Rheumatoid Arthritis [elevated above normal]
Hodgkins Lymphoma 215
Table 1.1 Adapted from Nutritional Biochemistry and Metabolism, with Clinical
Applications. Linder, M.C. et al., (1989) Arch. Biochem. Biophys.
In inflammatory disease, infection, and cancer, serum iron concentrations
decrease at the same time SFts levels increase. There are several plausible
explanations for this phenomenon. Studies have shown that administration of iron-
binding proteins, such as apotransferrin or lactoferrin, inhibits bacterial cell
proliferation. Conversely, the introduction of iron increases the virulence of bacterial
and other infections .
SFts concentrations are one of many methods utilized in the evaluation of
neoplastic disease. One type of analysis examines the so-called tumor antigenic
moieties. Immunohistochemistry involves the incubation of the suspected tumor
tissue with a well-characterized primary antibody directed at the antigen of interest.
A secondary (biotin-labeled) antibody is applied that is specific for and that binds the
first antibody. Subsequent binding of an avidin-biotin-horseradish peroxidase
complex, as well as a reduction/oxidation reaction in the presence of dye, allows for
microscopic visualization of the antigen. Some of the common antigenic moieties
include, CA-125 which is used to identify female genital carcinomas; thyroglobulin,
which is used to identify thyroid carcinomas; leukocyte common antigen, which is
used to distinguish between lymphomas, carcinomas, and melanomas; and prostate
specific antigen (PSA), which is used to identify metastatic prostate carcinomas.
Unfortunately, the antibody determinants to which these antigens are exposed are
rarely tissue or tumor-specific, meaning that the field is often narrowed but not
Tumor markers are similarly non-specific. Their presence in the normal
physiological condition is possible, and may appear in elevated levels with a variety
of neoplasms. The benefit of tumor markers, in comparison to the antigenic moieties
discussed above, is that they can be measured as part of a general blood panel in the
presymptomatic individual, whereas the antigenic moieties, are only measured after
the appearance of symptoms. There are five types of tumor markers. Hormones in
elevated concentration, such as calcitonin, suggest medullary thyroid cancer.
Oncofetal antigens such as alpha-fetoprotein are present with hepatocellular
carcinoma and various germ cell tumors. Isoenzymes are a class of tumor markers
that are a valuable aid in the detection of prostate carcinomas and small-cell lung
cancers. Specific protein markers are those that are tumor derived, such as PSA, and
are used in conjunction with host-response markers such as SFts, to aid in diagnosis
and staging . Because of the non-specificity of most of the tumor markers
available today, their usefulness is limited to screening for residual disease and
treatment response. Dr. Stuart Gordon of the University of Colorado Health Sciences
Center has developed a screening assay for a proteolytic enzyme produced by cancer
cells, but, because it is nonspecific in its diagnostic capabilities, the assay has finite
applications. If the specificity of these tumor markers used by the assays could be
enhanced, their clinical value would increase dramatically.
Other researchers have examined a variety of SFts regulation factors and
made connections to concurrent disease states or shown these factors to foster a
predisposition for the development of a disease state. SFts are negatively regulated
by the adenovirus oncogenes jun and fos. This suggests a protective role for SFts and
the ability of certain oncogenes to overwhelm this protection . Tumor necrosis
factor, on the other hand, stimulates the production of ferritin . This points to an
earlier step in the process when the serum ferritin is still operating in a protective
fashion to sequester free iron and deny developing neoplasms the nourishment they
require. Taniguchi reported that the presence of high concentrations of glycated
proteins, such as SFts, implies in vivo DNA damage . These complicated
regulatory mechanisms may converge and produce a scenario in which the protective
function of serum ferritin is overwhelmed and tumor cells can grow and divide
In a study reported in Cancer in 1989 Ji Ya-You, found that SFts levels
showed an inverse proportionality to the differentiation of cells in patients suffering
from malignant histiocytosis, a rare form of soft-tissue sarcoma. Ji found that serum
ferritin behaves like a tumor-associated antigen and, as the cell types progress from
well-differentiated to anaplastic, the concentration of serum ferritin increases .
Jis study may allow for diagnosticians to distinguish between benign and malignant
varieties of this sarcoma and permit the clinician to recommend therapy based on the
measured SFts value. A later study in the same journal reported a prognostic
connection between SFts levels and staging in 50 children diagnosed with lymphoma.
In this study, Harm and Lange looked at the correlation between SFts levels and
advancing disease stage of childhood Hodgkins Disease. As SFts levels increased
above the normal (normal = 7 to 142 ng/mL for children older than 6 months), the
four-year survival of the patients opposite correlative, transferrin levels, also proved
to be a valuable, but the prognostic outlook for the patient was poor . Erythrocyte
sedimentation rates as well as hemoglobin and SFts levels were analyzed and, as the
SFts levels increased, there was a corresponding linear increase in the staging of the
disease (tumor level classification) and a decrease in the progression-free survival
The figure shown illustrates the connection between protein concentration and
Figure 1.4 Adapted from Hie-Won L. Hann, et al, Cancer 7:15-90. Graph shows
serum ferritin levels versus cancer stage in Hodgkins disease.
Like the Ji study, this research may permit the clinician to more accurately
stage the lymphoma and determine the appropriate treatment more quickly.
2. Problem Statement
2.1 The Clinical Use of the Ferritin Assay
The medical community has assayed levels of ferritin in the serum of patients
for many years. This suggests that SFts may be significantly more important, on a
clinical level, than the cellular and glycosylated heart ferritins described above.
Studies done on the serum of healthy horses, as well as other animals, have
established that trace amounts of SFts are present (approximately 100 ng/mL) in the
normal individual. Having witnessed this in the animal model, the medical
community began employing the serum ferritin assay as a diagnostic tool. As
previously mentioned, when a disease state exists, the level of SFts increases
dramatically as the disease progresses.
In the past fifteen years, over 700 studies have reported SFts concentration
increases in relation to various diseases based on commercially available ferritin
assay kits. Ferritin immunoassay kits employ primary anti-ferritin antibodies raised
against cellular ferritin, not serum ferritin. An early study by Richard Sheehan
analyzed a packaged kit assay manufactured by Ramco Laboratories, Inc. Sheehan
describes the pure ferritin standards from human tissues used by this particular
assay to estimate serum ferritin values. He goes on to acknowledge that,
Values however, cannot be considered absolute, since the binding
characteristics of human serum ferritin and the standards prepared from human tissue
sources are divergent above approximately 2 ng/pg (a calculated serum value of
approximately 40 ng/mL) .
Sheehans discussion of the inherent problems of this assay does not dissuade
him from describing the assay as relevant and valid. Journal articles by Halliday,
Luxton, Ryan, Anaokar, Icobello, and Revenant all refer to existing commercial
assays or propose new assays that use primary antibodies grown against the cellular
isoferritin (rabbits were injected with either human liver or human spleen ferritin)
[39,40, 41, 42,43,44]. The resulting polyclonal antibody described or suggested in
this assay is a mixture of antisera to the cellular antigen as well as a host of other
antibodies produced and contributed by the host animal. While SFts assayed by these
polyclonal antibodies have sufficient determinants to be recognized by the cellular
antibody, the mixed antibody does not permit quantitation of only CFt or only SFt
levels if that is desired to determine disease progression. The assays currently
available do not make this distinction and suggest the data obtained by using the
assay are valid SFts levels. The values reported by these kits for SFts, are, in fact, a
combination of SFt and CFt that may be present due to the tissue damage that
accompanies inflammatory diseases and invasive cancers. They might also be due to
the over-synthesis of the protein for purposes of iron sequestration, or to any of the
factors mentioned above. The mixed antibodies are not useful for the correlation of
specific SFt or CFt levels to disease stage.
It is critically important, therefore, that an appropriate assay be developed for
hospital use that can separate the two types of ferritin with clinical significance (CFt
and SFts) and then quantify the amount of CFt and SFts present in blood serum.
2.2 The Study and Assay Development
In this study human SFt and hemochromatosis SFt were purified according to
established protocol (see chapter 3.1). The purified human SFt protein was then used
to produce an anti-human SFt antibody, using a New Zealand White rabbit as a host.
The resulting mixed antibody was incubated with human cellular ferritin in order to
bind the cellular antibodies present in the mix, leaving the majority of the serum
antibodies free. The mixture was passed over a size exclusion column from which
the heavier cellular ferritin/cellular antibody complex eluted first. The column
fractions were analyzed spectrophotometrically and plotted. Samples from the
distinct peaks were used as primary antibodies in western blot electrotransfer and
showed that the purified antisera specifically recognizes serum antigen and not the
cellular variety. We now propose an assay to accomplish the goal of the separation
and quantitation of cellular and serum ferritin concentrations. This new approach to
utilizing ferritin values may prove to be a valuable gauge of both disease stage and
3. Materials and Methods
3.1 Purification of Human Serum Ferritin
Human serum ferritin was purified from 12 Units (3.25 L) of expired,
recovered human serum obtained from Belle Bonfils Memorial Blood Center,
Denver, Colorado. Serum was heated to 70C for 10 minutes, then placed on an ice
bath for 15 minutes to precipitate non-ferritin proteins. A volume of 1100 mL was
decanted and frozen for later processing. The remaining volume, 2500 mL, was
acidified with acetic acid (~5.5mL) to pH 4.8, followed by centrifugation at 9000 x g
for 10 minutes. The supernatant was retained and the brown, gelatinous pellet was
discarded. The heat treatment and centrifugation was then repeated once. The
supernatant was neutralized to pH 7 with IN NaOH bringing the new volume to 2150
mL. This volume was ammonium precipitated with 3.06 g ammonium sulfate per 10
mL volume and stored at 4C for 24 hours. Solution was then centrifuged at 10,000 x
g for 20 minutes. The supernatant was discarded and the pellet was retained.
This pellet was dissolved in 0.01M sodium phosphate (Ab) buffer, pH 7.4,
containing .15M NaCl, and 0.02% NaN3. The resolublized pellet was dialyzed versus
3 x 800 mL phosphate buffer (Spectrapore Membrane Tubing #1., mw cutoff 6000-
8000). The resolubilized pellet was loaded onto a 2 cm x 60 cm sephadex-G-150
column. Fractions in the amount of 0.5 mL were collected and analyzed at 280 nm.
The serum ferritin was further purified and concentrated by immunoaffmity
chromatography (see section 3.4)
3.2 Concanavalin-A Chromatography
In preparation for binding to the lectin Concanavalin-A (which binds high
mannose and complex carbohydrate side chains), dialyzed ferritin samples (see
section 3.1) were dialyzed again against Concanavalin-A buffer: 0.02MTris-HCl
buffer pH 7.4 containing 0.15M NaCl, ImMMn CI2, 1 mMCaCl2, ImM MgCh 0.02%
NaN3, and 0.2 mM PMSF. The ferritin was dialyzed in 1.0L of Concanavalin-A
buffer with repeated buffer changes for 24 hours. Concanavalin-A sepharose (Sigma
Chemical Co.) was packed in a 5 mL, column and 7 volumes of Concanavalin-A
buffer were run over the column to remove residual preservatives. Ferritin samples
were then diluted in 6.0 mL Concanavalin-A buffer and applied to the column. The
ferritin sample was chased with 15 mL of Concanavalin-A buffer, and 1.0 mL
fractions were collected. A volume of 10 mL of buffer containing 0.1M methyl alpha-
D glucoside was then applied to the column, followed by 10 mL of buffer containing
0.5 M methyl alpha-D glucoside. 1.0 mL fractions were collected. Finally, 25 mL of
Concanavalin-A buffer containing 0.5 M methyl alpha-D mannopyranoside was
applied to elute glycosylated proteins from the Concanavalin A-sepharose column.
Eluted proteins were monitored spectrophotometrically at 280 nm and by SDS-
3.3 Protein Assay
A well-mixed solution containing 0.200 mL of 10% dye reagent (diluted in
H2O) was added to 0.800 mL of protein sample. The samples were vortexed and
allowed to stand for 5 to 60 minutes to allow for color development. Absorbency was
read at 595 nm. Bovine serum albumin was used as a standard.
3.4 Anti-Ferritin Immunoaffinity Chromatography
Human placental ferritin antibodies (supplied by Dr. Paul Seligman UCHSC)
were covalently bound to Sephadex gel (2 mL volume) by CNBr activation. The
ImmunoPure Antigen/Antibody Immobilization kit (Pierce Chemical Co.,
Rockville, IL) was utilized in construction of the column. Volumes of 40 to 50 mL of
purified human serum ferritin (see section 3.1) were applied to the immunoaffinity
column and recirculated once. The unbound protein was then washed with 20 mL of
Ab column buffer (see below) and 1 mL fractions were collected. Fractions were
detected spectrophotometrically. Specifically bound proteins were eluted by the
addition of 8 mL of 0.1 M glycine, pH 2.8, and collected in 1 mL fractions. The
column was washed with 20 mL phosphate buffer and stored in phosphate buffer
containing 0.05% NaN3.
1.5 Preparation of Anti-Human SFt Antiserum
A 5 lb. female New Zealand White rabbit was housed at the University of
Colorado Health Sciences Center Animal Resource Center. Antiserum was prepared
iccording to protocol number 13514792(07)1 C. The injection and boosting
procedure was performed by a Resource Center technician in collaboration with the
rnthor and Dr. Paul Seligman. Serum ferritin (See section 3.1., 1.0 mL, O.lmg/mL)
Durified as described above, was combined with 0.5mL adjuvant (TiterMax, Hunter
he.). On day one, the rabbit, which had been acclimated for 7 days, was given two
mbcutaneous injections of the antigen/adjuvant complex. This was repeated on day
H. Next, two O.lmL subcutaneous injections of antigen (SFt) without adjuvant were
jiven in each shoulder on days 28 and 29. On day 31, 0.5 mL antigen injection was
jiven intraperitoneally. On day 36,10 mL of blood was obtained from the rabbit
hrough a capillary ear bleed. On day 40, approximately 60 mL of blood was
collected from a femoral artery. Resource Center personnel under supervision of the
Center Veterinarian euthanized the animal. Serum was separated by centrifugation at
1C at 2000 rpm in a clinical centrifuge and the sample was frozen for storage.
1.6 Sephacryl Gel Exclusion Chromatography
Hydrated Sephacryl S-200 (Sigma Chemical Co.) was poured into a
chromatography column and equilibrated by running 500 mL of NaHCC>3 buffer (pH
I. 5 obtained with IN NaOH). The void volume was calculated using 1.0 mL of
0.05% blue dextran prepared in the bicarbonate buffer. The elution position of
known molecular weight standards were established by application of the following
standards: cytochrome C (mw. approx. 12 kD), hemoglobin (mw.64,500 Da), CFt
(450,000 Da, Sigma Chemical Co.), phenol red (mw. 354 Da), and blue dextran (mw.
2,000,000 Da). A standard curve was prepared for extrapolation of a-SFt antibody
sample data. 1.0% of the column volume of antibody (previously subjected to a 50%
ammonium sulfate precipitation in an effort to bring down the antibody-antigen
complex) was applied, followed by bicarbonate buffer, and 1.0 mL fractions were
collected. Fractions were analyzed spectrophotometrically at 280 nm.
3.7 Polyacrylamide Gel Electrophoresis
SDS-PAGE was performed on an 8 x 8 cm, 0.75 mm thick minigel in a Life
Technologies Mini-V Vertical Gel Electrophoresis Apparatus, BRL, Gaithersburg,
MD). Separating gels were 15% acrylamide, pH 8.8 (30:0:8 acrylamide: bis-
acrylamide), and stacking gels were 4% acrylamide pH 6.8, as previously described.
Protein (1.0 mg) was solubilized in a 1:1 mixture of 0.125 M Tris-HCl buffer, pH 6.8.
The mixture contained 4% SDS, 20% glycerol, and 10% P-mercaptoethanol (2X
buffer) and was boiled for 3 minutes. Low molecular weight standards (BioRad,
Richmond, CA) were run simultaneously in the outside lanes. Electrophoresis was
conducted at 194 volts constant current for 40 to 50 minutes in an ice-water bath.
Running buffer containing 0.02M Tris, 0.192M glycine, and 0.1% SDS was utilized
during the procedure to fill the Life Technologies box. Mini gels were stained in
0.25% Coomassie Blue R-250, in 40% methanol and 7% acetic acid. Gels were
destained in 7% acetic acid and 5% methanol for 1 to 24 hours at 25C. Gels
intended for Western Blot electrotransfer (see section 3.9) were not stained until after
the electrotransfer. Molecular weights of separated subunits were assigned based on
the migration Rf position of the standards relative to the tracking dye.
3.8 Western Blot Electrotransfer
Proteins were separated by SDS-PAGE, as described above, and then
electrotransfered (BRL: Life Technologies transfer unit) onto nitrocellulose sheets at
194 volts for 40 minutes. Transfer was conducted at 15C in a buffer containing
25mM Tris-HCl, 0.14 M glycine and 20% methanol (Towbin buffer). Nitrocellulose
sheets were blocked for 24 hours in 20 mM Tris buffer containing 4% dry milk, pH
7.5. Blocked nitrocellulose sheets were washed in TBS buffer containing 0.05%
Tween-20 (TTBS) for 10 minutes with continuous agitation. The Nitrocellulose
sheets were incubated for 24 hours in a (1:100) dilution of anti-human serum ferritin
antiserum or a (1:100) dilution of anti-human spleen ferritin antiserum (Sigma
Chemical Co. St. Louis, MO). The nitrocellulose sheets were then washed twice in
TTBS for 10 minutes with agitation, and incubated for 24 hours in TTBS buffer
containing a 1:3000 dilution of alkaline phosphatase-conjugated goat anti-rabbit
antibody (Sigma Chemical Co.). The conjugate solution was decanted and the
nitrocellulose sheet was washed twice in TTBS for 5 minutes. The sheet was next
washed in TBS for 5 minutes to remove residual Tween-20. Developing buffer was
prepared by dissolving 30 mg para-nitro blue tetrazolium chloride (NBT) and 15 mg
of 5-bromo-4-chloro-3-indoyl phosphate para-toluidine salt (BCIP) separately in 1
mL 70% N,N-dimethylformamide, which was then added to 100 mL 0.1M NaHCC>3
pH 9.8 carbonate buffer with 1.0 mM MgC^. The final concentrations were 0.3
mg/mL NBT and 0.15 mg/mL BCIP. Color was developed by immersing the
nitrocellulose filters in the developing solution for 30 minutes. When adequate
contrast was observed, color development was halted by immersion in
3.9 Radiolabeling of SFt
Into a plastic test tube the following were added successively: 10 mL purified
SFt (1.1 mg/mL) made in 0.05 M sodium phosphate buffer, pH 7.4; 5 mL radioactive
iodide, 50 mL 0.5 M sodium phosphate buffer, pH 7.4; 2 mg/mL Chloramine-T in
0.01 M sodium phosphate buffer containing 1 M NaCl, 0.1% BSA, and 1% KI, pH
7.4. The solution was mixed gently and allowed to stand for 30 s for radioiodination.
500 mL of 1 mg/mL sodium metabisulfite in 0.01 M sodium phosphate buffer
containing 1 M NaCl, 0.1% BSA, and 1% KI, pH 7.4 was added and acted as a
reductant to stop radioiodination. The solution was then purified over Sephadex G-25
(see section 3.6) with the radioiodinated protein placed into the gel bed. Fractions
were collected and cpms counted on a Beckman 4000 gamma counter.
3.10 Double Antibody Radioimmunoassay
A. Ferritin samples were analyzed using the commercially available assay
produced by Diagnostic Products Corporation. A standard curve using provided
calibrators (0, 10, 25,100, 200, 500,and 1,000 ng/mL CFt) was established prior to
each assay. Duplicate ferritin samples were then run according to assay instructions:
100 mL of purified ferritin was placed into 5 mL polyurethane tubes. A volume of
100 mL of I Ferritin and lOOmL of Ferritin antiserum was added to the tubes. The
tubes were then vortexed and incubated for 1 hour at 37C. A solution 1.0 mL of cold
precipitating buffer (containing goat anti-rabbit gamma globulin and dilute
polyethylene glycol in saline) was added to the tubes, and they were then vortexed
prior to centrifugation at 3000 X g for 15 minutes. The supernatant was decanted and
the gamma decays of the pellet were counted for 1 minute.
4. Results and Discussion
4.1 Purification of SFt
The initial step in this study, involved purification of the human blood serum
with the goal of retaining the ferritin. Purification by precipitation is a common
method, and precipitation by denaturation lends itself to large initial volumes like
those used in this study. Precipitation by denaturation can be used as a purification
step if the protein of interest is not denatured by the treatments. Ferritin can
withstand temperatures up to 70C without the tertiary structure of the protein being
disrupted. When this disruption occurs, random coil structures are formed. These
coils become entangled with one another, and aggregates develop composed of the
contaminant proteins . Specifically, high temperatures induce denaturation by
breaking many of the bonds (van der Waals forces, ionic interactions, hydrogen
bonds) holding the protein in its native conformation. Different proteins are
denatured at different temperatures.
The second step of the purification also employed precipitation by
denaturation, but in this case, pH changes were utilized. Extremes of pH cause
internal electrostatic repulsion, or loss of internal electrostatic attraction by changing
the charges on the side chains of the amino acids. This technique opens up the
protein and the bound solvent is lost causing denaturation.
The goal of any purification is to obtain a maximum yield of protein with
maximal purity of the product . Overall, the choice of purification techniques and
the number of steps used should be well ordered and minimized. For each protein,
certain properties can be exploited to increase capacity: the amount of the sample in
terms of volume and concentration, resolution; the efficiency of separation from
contaminant proteins, and the protein yield. The method of choice, ammonium
sulfate precipitation, usually gives product yields better than 80%. In this study,
crude SFt was separated on sephadex G-150. Figure 4.1 shows the elution profile and
Figure 4.2 shows a poly acrylamide gel for the initial purification, as well as, later
more refined purifications .
Figure 4.1. Elution profile for initial serum purification. In this study, the initial
elution of the crude SFt was separated on sephadex G-150. Fractions 27 through 33
were pooled and analyzed by SDS-PAGE (see fig. 2). Comparison lanes containing
horse serum were run at the same time and confirm that fractions 27 through 33
contained bands typical for SFt with subunits of 66, 60 and 45 kD.
Lanes: 1 2 3 4 5 6 7
Figure 4.2 SDS-PAGE. Lane 1 LMW standards. Lane 2,3 HoSFt initial purifications,
Lane 4 blank, Lanes 5,6,7 SFt.
Immunopurification is a highly selective and powerful purification method
because antibodies can be used that distinguish between very similar antigens and
thereby overcome the difficulties encountered in other purification protocols.
Immunopurification is often a later step in the purification of proteins because
production of the necessary antibodies is relatively expensive and because, there is a
desire to protect the antibodies (which are valuable), from proteolytic attack or
fouling by contaminants in the crude protein solution . Some of the commercially
available immunoprecipitation kits require activation with cyanogen-bromide, a
dangerous, highly toxic chemical. Some of the newer kits employ preactivated
matrices that can simply be mixed with the antibody in a suitable buffer, then washed
and blocked following an established protocol. The key step involves the selection of
the elutant. Most antibodies can withstand extremes of pH and ionic strength, so the
nature of the antigen dictates the composition of the elutant. Elution is obtained by
breaking the bonds, such as salt bridges, hydrogen bonds, or Van der Waals forces
which form the peptide-peptide complexes. The elution is usually accomplished by
changing the ionic strength, pH, or temperature. In this study, pH was altered.
Polyclonal antibodies have much less molecular recognition than do monoclonal
antibodies. As a result, the polyclonal ab may have multiple cross-reactions with the
antigen, thus reducing the specificity. Conversely, two protein antigens (the labeled
and non-labeled proteins) may share extremely small, precise details of their surface
topology which permits bindings not available to conventional ab and antigens.
Monoclonal ab (MAbs) do not guarantee molecular recognition because two antigens
having dissimilar surface structure may have similar epitopes that can interact with
the same ab. The MAbs, however, when made with care against pure antigens,
significantly improve the antigen/antibody binding .
Figure 4.3 shows an SDS-PAGE comparing preparations of CFt from horse,
SFt from horse, as well as SFt purified in this study. The comparison is interesting in
that slight differences can be seen in the molecular weights, and banding patterns.
Lanes 1 2 3 4 5 6
Figure 4.3 SDS-PAGE of SFt purification, lane 1, LMW standards, lane 2, HoCFt.
lane 3,HoSFt initial purification, lanes 4,SFt initial purification^ SFt refined, lane 6,
Concanavalin-A lectin (MW. 55,000) from the jack bean binds glycoproteins
with high mannose or complex carbohydrate side chains. The affinity of lectins for
specific carbohydrate moieties makes them particularly useful for purifying distinct
groups of glycoproteins. For a lectin such as Concanavalin-A to function as a
purification media, it must be coupled to a solid-phase, such as Sephadex G-100, and
then immobilized. In this study, MnCL and CaCL were utilized as divalent metal
ions. These ions result in a transformation in the three-dimensional structure of the
lectin that ultimately leads to the formation of the saccharide binding site. By
including the appropriate metal ions, prior to coupling, it is possible to ensure that the
lectin is in the correct form to bind the protein . This allows both specific binding
of any glycoprotein containing a-D mannopyranosyl or a-D glucopyranosyl residue
with the C-3, C-4, and C-6 hydroxyl groups, and the subsequent dissociation using
selected elution conditions. Typically, 20 mM glycine-HCl at pH 2.0, along with 0.1
M a-methyl mannoside in PBS (poly-buffered saline), is used to elute the product.
Binding of the glycoprotein is pH-dependent; therefore, samples are buffered at
physiological pH. The advantage of lectin chromatography is the ability to run the
column in the presence of high salt concentrations. Binding between the sugar side
chain and the lectin is not due to ionic interactions, and the salt prevents non-specific
binding of proteins within the matrix. SFt has the sugar side chains, whereas CFt,
does not contain the sugar side chains, and therefore does not bind to the
Concanavalin-A matrix. This has proven to be a good technique for isolating the SFt
from other ferritin types. Figure 4.3 shows the elution profile for the purification of
the protein using the concanavalin-A lectin.
With the peptide at its best attainable level of purity and activity, the next goal
involved the production of an antibody to the SFt. Commercially available CFt
antibodies are commonly in use. a-SFt is not available.
Figure 4.4 Elution profile for Concanavalin-A purification of SFt. The diagram
shows that with the addition of the 0.5 M glucoside, the ferritin antigen is released.
SDS-PAGE confirmed the presence of SFt in the second peak.
An antigen, by definition, stimulates the production of antibodies, which in
turn combine with the antigen. The stimulation and combination are based on
complementarity or fit between the two shapes. A small piece of the antigen and
the combining site of the antibody link, and an antibody/antigen complex is born .
The closer the fit between this site and the antigen, the stronger will be the non-
covalent forces (hydrophobic, electrostatic, etc.) between them, and the higher the
affinity . It is with this objective in mind, that the antibody is prepared.
Specificity and affinity ultimately determine the usefulness of an assay employing
The immunization procedure may be as important as the antigen in
determining success in developing antibodies . It is desirable to stimulate only the
B lymphocyte cells that will make high-affinity antibodies. The current hypothesis is
that the immunoglobulin antigen receptor site is expressed at the surface of the
lymphocyte. That suggests, therefore, that only enough antigen should be provided
during the immunization period to be absorbed by those cells with high affinity sites
available . The dosage of immunogen (antigen) appears to modulate the class of
antibody formed. In addition, the capacity of the host animal to produce antibodies
determines whether IgM, IgG, IgA, or one of the other classes of antisera is formed
High affinity, highly specific antibodies can be formed to single antigens
using mice as the host animal. These so-called monoclonal antibodies increase the
antigen-antibody complex that forms at equilibrium but the production costs are
prohibitive and as such, the monoclonal variety are less frequently used in assays.
For this study, a polyclonal antibody was produced. After following a conventional
immunization protocol, the animal was exsanguinated and the immunoglobulins were
purified. The properties of an antiserum that must be assessed are the titer,
sensitivity, and specificity. Optimal specificity is obtained when approximately 50%
of the labeled antigen are bound by antibody in the presence of unlabeled antigen
. Specificity is often used to mean that the antiserum does not cross-react with
other proteins likely to be encountered in the samples. While specificity can be
calculated mathematically, in this study the relative specificity was assessed by
western blot electrophoresis. The competitive specificity described above was
examined within the confines of the radioimmunoassay. A Sephacryl column was
used to further purify the antiserum. The column was loaded with blue dextran,
phenol red, and antiserum previously treated with a 50% ammonium sulfate
precipitation to bring down the desired antigen/antibody complex.
Fractions were collected in 500 mL volumes and analyzed
spectrophotometrically at 280 nm. (Figure 4.4)
0 20 40 60 80 100 120 140 160 160 200
Figure. 4.5 Elution profile for a-SFt. Both cpms (counts per minute) and absorbance
at 280 nm are shown. Peak I, fractions 45-61, contained antiserum aggregates. Peak
II, fractions 73-110, contained the purified a-SFt.
The Concanavalin-A fractions were pooled and the purity and specificity of
the antiserum was further confirmed by western blot electrophoresis.
The migration of proteins in polyacrylamide gels in the presence of sodium
dodecyl sulfate (SDS) is a function of their molecular weight depending upon the
pore size of the gel matrix. The pore size is a function of the degree of cross-linking
between the acrylamide molecules which is proportional to the acrylamide and bis-
acrylamide concentrations. The size of the pore is also important at the time of
eluting the proteins from the gel. In essence, the larger the protein, the slower the
transfer; the larger the pore size, the faster the transfer . In this study, Western
Blot was utilized for example, to compare a-CFt, and subsequent specificity of and
a-SFt. Figure 4.5 is an example of such an experiment. What is notable about this
western blot is the high specificity of the a-SFt for the SFt antigens. Similarly, we
see no cross reactivity between the a-CFt and the SFt antigens; a positive result.
Figure 4.6 Western Blot Electrotransfer. The top frame shows anti-serum ferritin (a-
SFt) versus the following antigens; lane 1 LMW standards, lane 2 purified SFt, lane 3
blank, lane 4 blank, lane 5 SFt (purified from a patient with leukemia), lane 6 blank,
lane 7,blank 1 Jane 8 blank, lane 9 CFt (from a patient with hemochromatosis), lane
10 LMW standards. The lower frame has identical lanes but was run versus anti-CFt
4.2 Radiolabeling of 125I to ferritin
Successful radioiodination is critical for the satisfactory performance of an
RIA (radioimmunoassay). The so-called Chloramine-T method is the best known of
the many available methods of iodination . The Chloramine-T method was first
described by Greenwood and co-workers in 1963 and remains essentially unchanged
since then . In this method, an aqueous solution of chloramine-T (the sodium salt
of the N-mono-chloro derivative of p-toluene sulfonamide) breaks down slowly to
form hypochlorous acid. Sodium iodide is then oxidized under mildly alkaline
conditions (pH 7.5) by the hypochlorous acid, and cationic iodine is formed. With
the tyrosine residue as the target, the iodination proceeds and the iodine atom
substitutes in the ortho position to the hydroxyl group in the phenol ring (see fig.4.7
below). Some proteins being labeled exhibit substitution at different tyrosine residues
depending upon varying degrees of accessibility .
The initial RIA assay used the purified SFt and lab produced a-SFt. The assay
confirmed the viability of the assay in that the majority of the CPM came down in the
pellet as would be expected. There was a single positive result, but attempts to
duplicate the experiment have been unsuccessful. This might be attributed to
inadequate binding of the antibody-antigen complex, problems with the PEG
precipitation, destruction of binding sites, or a host of other problems. In a process
with multiple steps each one must be explored and determined to be individually
competent as well as working in concert in order for the complicated assay to
Figure 4.7 Diagram of Monoiodotyrosine, showing the ortho location of the iodine
after successful iodination.
There have been a variety of problems identified with the chloramine-T
method. Variations to the method have been incorporated over the years to overcome
the many of the main difficulties. The oxidizing action of chloramine-T may damage
some proteins and render the product unstable; therefore, the concentration of the
reagent and the reaction time are reduced as much as possible. The standard protocol
reagent and the reaction time are reduced as much as possible. The standard protocol
calls for 50 mg of chloramine-T per 10 mL of protein to be labeled. The chloramine-
T ratio was reduced to 16 mg per 10 mL of protein and we obtained satisfactory
results. The amount of radioactive reagent is also kept to a minimum to avoid the
potential incorporation of too much iodine, which can in turn lead to poorer binding
and recognition. Sodium metabisulfite, the reducing reagent, used to stop the reaction,
is often suspected of causing damage to the protein. There is evidence that the
bisulfite may cleave the internal disulfide bonds. The protocol calls for 200 mL
sodium metabisulfite per 200 mg protein along with 100 mL per 1 mg protein of
potassium iodide. Some researchers have obtained better results by reducing the
sodium metabisulfite concentration to 6 mL per 10 mg protein. In this study the
standard protocol was utilized. The most important property of the iodinated product
is its performance in the radioimmunoassay.
After iodination, purification by gel filtration or ion exchange chromatography
is used to remove excess iodine. In this study, Sephadex G-25 was used. In this case,
fractions 24-38 were pooled, and frozen at -70C in aliquots. Figure 4.7 shows the
elution profile for the purification of the labeled I ferritin. The second peak was
the peak of interest, because it contained the labeled protein.
Label 1JSI and Ferritin
Figure. 4.8 Elution profile of 125-1, SFt purification. Peak I shows the void volume
(note Peak I contains fractions25-30), peak II shows the labeled protein, and peak III
shows the free iodine.
4.3 Comparison with commercial antigens and kits
Kits are provided with ab that have been grown against CFt in the form of
spleen ferritin. In these commercial kits,1251 labeled ferritin competes with ferritin in
the patient sample for sites on ferritin specific antibodies. After incubation,
separation of bound from free ferritin is achieved by a PEG-antibody method
identical to the one utilized in this study and described in sections 3.9 and 3.10. The
tubes are counted on a gamma counter, the counts being inversely proportional to the
amount of ferritin present in the patient sample. The kits, therefore, are not highly
specific. They do not permit the researcher to easily separate and quantify SFt and
CFt. This step, and the development of a working RIA could prove important in the
management of anemias and some malignancies.
The next series of Figures 4.9 through 4.11, show SDS-PAGE of the
calibrators provided by the commercial kits for standardizing patient samples. It is
apparent from these PAGE results that the calibrators are composed of highly impure
samples, and, while one can extrapolate the concentration of total SFts present, it
would not be possible to differentiate between SFt and CFt. Following Figure 4.11,
there are series of figures (Figures 4.12 and 4.13) showing the reactivity of the
commercial kit calibrators to the antibody produced in this study, as well as a-Cft to
the antigens supplied by the commercial kits.
Figures 4.9,4.10,4.11 SDS-PAGE of calibration standards supplied in commercial
serum ferritin analysis kits. Three kits were compared and all showed multiple
protein bands in addition to the stated SFts bands. All three kits showed evidence of
protein bands corresponding to the CFt molecular subunit weights of 19kD and 21kD.
Figure 4.12, 4.13 The western blot electrotransfers of the three commercial kits show
ferritin calibrators versus a-SFt in figure 4.12 and versus a-Cfit in Figure 4.13. Lane 1
in both figures is LMW standards. Lane 2, Diagnostic Products 20 ng/mL), Lane 3
Ramco Industries-20 ng/mL. Lane 4, Abbott Corp 20 ng/mL. Figure 4.13 is as
follows: lane 2, Blank, lane 3 Diagnostic Products 20 ng/mL), Lane 4 Ramco
Industries-20 ng/mL. Lane 5, Abbott Corp 20 ng/mL. All concentrations reported
are for ferritin calibrators supplied by the various kits. Calibrators are supplied in
protein buffers. Proprietary limitations prohibit the commercial suppliers from
revealing the makeup of those buffers supplied in protein buffers.
What is the clinical significance of serum ferritin levels in human blood? It is
evident from this study that ferritin is present in the blood in very low concentrations:
usually less than 1% of the iron found in serum is stored in the ferritin molecule .
It is known that the serum ferritin is in equilibrium with body stores and that any
fluctuations in the storage of iron in the body are reflected in the ferritin
concentration. Serum ferritin levels decline very early in the development of iron
deficiency. Because this change is evident much earlier than observable changes in
blood hemoglobin concentrations, red blood cell size, or serum iron levels, an
argument could be made that measurement of SFt would allow for earlier diagnosis
and lead to more effective treatment of various anemias. In the early stages of iron
deficiency, there is less likelihood that other concurrent diseases will be present .
Many other chronic diseases result in elevated serum ferritin levels. In instances of
chronic infection, chronic inflammation, and various malignancies, SFt levels can be
used as a nonspecific indicator that informs the clinician to investigate further. When
any of the disorders mentioned above occur simultaneously with iron deficiency,
which is frequently the case, SFt levels tend to stay in the normal range. It is in these
cases, that a highly specific assay, one that can aid in differentiating between SFt and
CFt levels should be available to aid the clinician in making an appropriate diagnosis.
Diagnosing anemias present concurrent with high SFt levels would not be possible
with current assay technology.
Chapter 5: Summary/Future Studies
While this remains a project in development, we can say that we have
accomplished the following goals: first, we successfully purified SFt from human
blood. This novel protein is relatively unstudied and holds significant clinical
promise. Its purity was confirmed by polyacrylamide gel electrophoresis. Next, we
compared the purity of the novel SFt protein that we prepared to previous published
studies. Using the purified SFt, we produced a novel polyclonal antibody in a rabbit.
Following standard purification protocols, we purified the antibody and then
confirmed its purity and specificity using established western blot electrotransfer
We began development of a practical assay using the study antibody (a-SFt)
and the purified antigen. We looked at CFt, as well as the commercially available SFt
kits. We found that our antibody showed minimal and often no cross reactivity with
CFt and that the commercial kits, while claiming to be diagnostic for SFt, have wide
cross reactivity with CFt and contain calibration standards that are significantly
impure. They can report values for SFt but cannot distinguish it from the other
isoferritins and cannot claim pure SFt values.
It is our contention that the study antibody (a-SFt), with it high specificity,
along with an antigen of high purity that could produce calibrators of high purity,
would allow the clinical investigator or laboratory personnel to distinguish CFt from
SFt, or other isoferritins, and to make important prognostic recommendations on
behalf of the patient.
Ultimately, the goal of this project is to improve the viability of antibody
production as the foundation of the assay. The rabbit as the production vehicle must
be eliminated. The monoclonal antibody should be the method of choice. MAbs are
produced as products of cloned lymphoblastic cell lines . MAbs have advantages
over polyclonal antibodies in that they have increased specificity, can be produced in
relatively large quantities of consistent quality, and the cell lines that secrete the
MAbs are essentially immortal. Some studies report problems with unexpected or
diverse cross-reactivity with Mabs . We anticipate, however, that with the
excellent results seen thus far with the polyclonal a-SFt antibody (in terms of
minimal cross reactivity with CFt), that the a-SFt prepared as a MAb will work well
in the assay.
Once the antibody can be produced consistently, and in bulk, automation
would be the next step. Accuracy, meaning being able to reproduce the correct result
(maximum cpm present in the pellet), would be critically important if the assay is to
have clinical feasibility.
Another issue needing to be resolved, is that of iodination. Proteins need to be
labeled, when l25I is used, every 6 to 8 weeks. In this study, the I25I labeled ferritin
consistently averaged >500,000 cpm immediately after purification. This is
obviously highly radiotoxic if handled incorrectly, and as a result, limits the use of
this assay to specially certified radiation hoods, and technicians. A future assay,
might employ a less radiotoxic method such as the IRMA (immunoradiometric assay)
method which immobilizes the MAb on the interior surface of a polystyrene tube
along with I25I tracer. The 125I still has the same half life, of approximately 60 days,
but the manageability of the assay is greatly increased.
Our goal is to maximize the affinity of the antigen/antibody complex, be able
to produce the antibody in bulk, insure the accuracy of the assay outcome, and
improve the precision of the assay. If all these goals can be achieved, the assay can
be developed to use SFt as a diagnostic and clinical tool.
BCIP 5-bromo-4-chloro-3-indoyl phosphate para toluidine salt
BSA bovine serum albumin
CFt cellular ferritin
cpms counts per minute
FMNH2 Flavin mononucleotide reduced form
HoSFt Horse Serum Ferritin
IRE BP iron responsive element binding protein
LMW Low molecular weight (refers to the standard markers used in
MABs Monoclonal Antibodies
mw molecular weight
mRNA messenger ribonucleic acid
NBT para-nitro blue tetrazolium chloride
rbc red blood cell
SDS sodium dodecyl sulfate
a-SFt Anti-Serum Ferritin (Human)
SFt serum ferritin
SFts serum ferritin(mixed isoferritins in the blood)
TBS -1 M Tris buffered saline
TTBS -1 M Tris buffered saline with 0.05% Tween -20
Wbc white blood cell
1. Brown, B. Hematology: Principles & Procedures 6th Edition. Lea & Febiger
Press, 1993, Philadelphia.
2. Brock, J.H. ed., Iron Metabolism in Health and Disease. WB Saunders Co.
LTD, 1994, London, pp. 17-32.
3. Brittenham, G.M. (19911 Hemotology: Basic Principles and Practice (eds. R.
Hoffman, R. Benz, S, Shatill, B. Furie, and H. Cohen), Churchill Livingstone,
New York, pp.327-349.
4. Halliday et al., (1990) Ann. NY Acad. ofSci. 101-112.
5. Brock, J.H. ed., Iron Metabolism in Health and Disease. WB Saunders Co.
LTD, 1994, London, pp. 102-104.
6. Dickey, L.F. Sreedharan, S., Theil, E.C., Didsbury, J.R., Wang, Y.H., and
Kaufman, R.E., (1987)7. Biol. Chem. 262,7901-7907.
7. Dietzel, J., Hirzmann, J., Preis, D., Symmons, P., Kunz, W.,(1992) Mol.
Biochem. Parisitol. 50,245-254.
8. Theil, E.C. (1987) Annu. Rev. Biochem 56: 289-315.
9. Worwood, M. (1986) Serum Ferritin. Clin Sci 70:215-220.
10. Munro, H.N., Linder, M.C. (1978) Physiol Rev. 58 317-396.
11. Harrison. P.M. (19871 Iron Transport in Microbes. Plants and Animals (eds.
G, Winkelmann, D. van der Helm) VCH Weinheim, pp. 445-475.
12. Linder, M.C., Goode, C.A., Gonzales, R., Gottschling, C., Gray, J., and Nagel,
G.M., ( 1989) Arch. Biochem. Biophys. 273, 34-41.
13. Moroz, C. Shterman, N., Kupfer, B., Ginzburg, I., (1989) Proc. Nat. Acad. Sci
USA 86, 3282-3285.
Murias, ML, et al., ( 1993) Mutation Research 295, 47-54.
15. Brock, J.H. ed. Iron Metabolism in Health and Disease. WB Saunders Co.
LTD, 1994, London, pp. 317-318.
16. Bradbear, R.A., Bain, C., Siskind, V. (1985) JNatl. Cancer Inst. 75, 81-84.
17. Britton, R.S., et al. Gastroenterology (1991)
18. Cragg, S.J., Wagstaff, M., Worwood, M.,(1981) Biochem J199, 565-571
19. Campbell, C.H., Crocker, D., Gruntmeir, J.J., Head, M., Kelly, T., Langfur,
M.I., Leimer, A.H., (1993) J Cell Biochem. 54, 1-13.
20. Zhou, Y. J. et al ( 1993) Mol Biol. Cell 3, 655.
21. Worwood, M. Cragg, S.J., Wagstaff, M., Jacobs, A., (1979) Clin Science 56,
22. Esumi, N., Ikushima, S., Hibi, S., Todo, S., Imashuku, S., (1988) Cancer
23. Kraus, JR, Stoic, V.,(1979) Am. J. Clin. Path. 26. 770.
24. Seligman, P.A., et al., (1987) in Molec. Basis of Blood Disease. Neinhuis, A.,
25. Lustbader, E.D., Hann, H.W., Blumberg, B.S., (1983) Science 220,423-425.
26. Schulof, R.S., Bockman., R, Garofalo, J., Cirrincione, C., Cunningham-
Riindels, S., Fernandez, G., Day, N., Pinsky, C., Incefy, G., Thaler, H., Good,
R., Gupta G., (1981) Cancer 48, 964-73.
27. Fakeir, A., (1991) (Meeting Abstract) J Tumor Marker Oncology 6(4) 57.
28. Williams, M.R., Turkes, A., Pearson, D., Griffiths, K., Blarney, R.W. (1990)
EurJofSurg Oncology 16,22-27.
29. Nishiya, K., Matsueda, H., Shirakami, T., Hatano, M., Yano, K., Ogura, T.,
Takaoka, M., Hraki, Y., Aona, K., Ezauni, K. (1985) Acta Med. Okayama
Hann, H.W., Evans, A., Seigel, S.E., Wong, K., Sather, H., Dalton, A.,
Hammond, D., Seeger, R. (1985) Cancer Res. 45, 2843-2848.
31. Kinijn, AM, Hershko, C. ( 1981) Br. JHaemotol. 49, 361.
32. Lasne, Y., Francina, A., Benzrara, O. (1991) Clin. Chim Acta. 197, 85-94
33. Campbell, C.H., and Linder, M.C. (1981) Biochim Biophys Acta 678:27.
34. Linder, M.C., Nutritional Biochem and Metab; Clinical Applications Second
Edition. Elsevier Press, 1991, New York, pp. 223-225.
35. Taniguchi, N., Kaneto, H., Islam, K.N., Hoshi, S., Myint, T. (1995)
Contributions to Nephrology 112, 18-23.
36. Ji Ya-You, et al., (1989) Cancer 64,1281-1289.
37. Hann, H.W.L., Lange, B., Stahlhut, M.W., McGlynn, K.A.. (1990) Cancer
38. Sheehan, RG., Newton, M.J., Frenkel, E.P., Am. J Clin. Path. (1978) 70, 79-
39. Halliday, J.W., et al., Clinica ChemicaActa (1975) 58,207-214.
40. Luxton, A.W., Walker, W., Gauldie, J., Ali, A.M., Pelletier, C. Clin. Chem.
41. Ryan, S., Watson, L.R., Tavassoli, M., Green, R., Crosby, W.H., Am. J.
Hematology (1978) 4, 375-86.
42. Anaokar, S., Garry, P.J., Standefer, J.C., Clin. Chem. (1979) 25,1426-31.
43. Iacobello, C., Ghielmi, S., Belloli, S., Arosio, P. Albertini, A., et al. Clin
44. Revenant, MC., Goudable, J., Beaudonnet, A., Mailliauin, A., Pichot, J.,
Monnet, L., Clin. Chem. (1985) 31,640-42.
45. Klausner, R.D., Rouault, T.A., Harford, J.B. (1993) Cell Press 72,19-27.
46. Findley, S., Podolsky, D., Silberner, J., (1992)US News & World Report 9-21,
47. Kerr, MA, ed., (1994) Immunochemistry. Bios. Scientific Publishers, p.70.