UNIVERSITY OF CINCINNATI
January 3, 2002
I, Arthur M. Barrie, III, hereby submit this as part of the requirements for the degree of:
Doctorate of Philosophy (Ph.D.) in: the Graduate Program in Molecular and Developmental Biology It is entitled: " The Role and Regulation of the Anti-Inflammatory Mouse Apolipoprotein J Gene"
Approved by: Dr. Bruce Aronow, Ph.D. Dr. John Bissler, M.D. Dr. Robert Colbert, M.D., Ph.D. Dr. Sandra Degen, Ph.D Dr. Jun Ma, Ph.D.
THE ROLE AND REGULATION OF THE ANTI-INFLAMMATORY MOUSE APOLIPOPROTEIN J GENE
A dissertation submitted to the
Division of Research and Advanced Studies of the University of Cincinnati
in partial fulfillment of the requirements for the degree of
DOCTORATE OF PHILOSOPHY (Ph.D.)
in the Graduate Program in Molecular and Developmental Biology of the College of Medicine
2001
by
Arthur M. Barrie, III
B.S., The Ohio State University, 1995
Committee chair: Dr. Bruce Aronow, Ph.D.
Abstract
Apolipoprotein J (apoJ) is a secreted protein with chaperone homology found in
vertebrates and is induced and deposited in both physiologic and pathologic states. ApoJ is
hypothesized to be a novel extracellular heat shock protein. I characterized the stress response
regulation of the mouse apoJ gene. Cell culture experiments verified the essential roles of an
apoJ promoter heat shock element (HSE) and the heat shock transcription factor 1 (HSF1) in heat shock-dependent apoJ gene activation. My studies identified multiple cooperating cis-elements, including a promoter activator protein-1 (AP-1) binding motif, an intron 1 enhancer, and a downstream silencer, that together conferred the full range of apoJ gene expression. I also sought to determine the function of the mouse apoJ gene in response to immune complex- mediated glomerulonephritis. I observed that apoJ deficiency exacerbated renal inflammation and predisposed to rapidly progressive disease. My data also revealed that apoJ levels affected the differentiation of the immune response. These experiments establish apoJ as a highly evolved heat shock protein that suppresses inflammation and immune-mediated tissue injury.
Dedication
I dedicate this thesis with love to my wife, Amy Felice, who has blessed me with endless
love and support during my trials and tribulations as a graduate student. I am very appreciative of the significant sacrifices she has made out of respect for my personal goals. I also dedicate this thesis to my parents, Sally and Boyd Frazier, and Arthur Barrie, Jr. My achievements are directly attributable to their love, patience, wisdom, and guidance, and for that I am eternally grateful. Lastly, I would like to humbly acknowledge Our Almighty Creator, whose handiwork and intellect are revealed in every scientific endeavor. Table of Contents
pages
General Introduction 5-26
Chapter 1
Cooperation between Multiple Cis-Elements of the
Mouse Apolipoprotein J Locus Is Required for Inducible
Gene Expression 27-102
Chapter 2
Suppression of Immune Complex-Mediated Glomerulonephritis
by Mouse Apolipoprotein J 103-160
Final Discussion
The Heat Shock Response and Mouse Apolipoprotein J 161-164
Bibliography 165-199
1 Figures and Tables
page
General Introduction
1. Map of the mouse apoJ gene and protein coding sequences. 22
2. Postulated structure of the mammalian apoJ protein. 24
3. Regulatory factors of the mammalian apoJ gene promoter. 26
Chapter 1
1. Schematic diagram of mouse apoJ gene sequences and reporter constructs. 66 Table 1. Wild-type and mutant mouse apoJ gene promoter element
sequences. 68 2. Basal expression profiles of transfected mouse apoJ gene wild-type and mutant reporter constructs. 70 3. Northern blot analysis of apoJ gene induction in mink lung
epithelial cells (CCL-64). 72 4. TGFβ responsiveness of transient transfected wild-type and
mutant mouse apoJ gene reporter constructs in CCL64 cells. 74 5. Northern blot analysis of apoJ gene induction in mouse
fibroblasts (NIH-3T3). 76 6. Heat shock inducibility of transient transfected wild-type and
mutant mouse apoJ gene reporter constructs in NIH-3T3 cells. 78 7. Time course study of transient transfected wild-type and
mutant mouse apoJ reporter gene expression during heat shock. 80 8. HSF1-transactivation of transient transfected wild-type and
mutant mouse apoJ gene reporter constructs in NIH-3T3 cells. 82 9. Heat shock induction of the mouse apoJ gene CAT-1 WT reporter construct in HSF1-deficient and wild-type mouse fibroblasts. 84
2 10. RT-PCR analysis of mouse apoJ gene heat shock induction
in HSF1-deficient and wild-type mouse fibroblasts. 86
11. Role of intron 1 enhancer during stress induction. 88 12. Role of downstream apoJ gene silencer on stress-mediated
induction. 90 13. Transgenic analysis of mouse apoJ reporter genes in response
to heat shock. 92
14. Reporter gene dissection of mouse apoJ gene intron 6. 94 15. Characterization of mouse apoJ reporter gene with partial
intron 6 deletion. 96
16. Promoter-mediated induction of the mouse apoJ gene. 98 17. Computer sequence analysis of the mouse apoJ gene intron 6
silencer. 100 18. Model of mouse apoJ gene regulation mediated by multiple
putative cis-elements. 102
Chapter 2
1. Scoring nomenclature for glomerular histopathology. 136
Table 1. Histopathological assessment of renal sections. 138
Table 2. Physiologic assessment of apoferrtin-treated mice. 140 2. Immunohistochemistry for IgG, IgM, and IgA in representative
glomeruli. 142 3. Immunohistochemistry for complement proteins C3 and C9 in
representative glomeruli. 144 4. Electron microscopic analysis of immune complexes in
representative glomeruli. 146
5. Characteristic electron microscopic glomerular images. 148 6. Serum complement C3 levels in wild-type and apoJ-deficient
treated mice in trials 1 and 3. 150
7. Anti-apoferritin IgG1 and total IgG1 serum antibody levels. 152
3 8. Anti-apoferritin IgG2a and total IgG2a serum antibody levels. 154
9. Ratios of Th2 versus Th1 serum antibodies. 156 10. Postulated mechanisms of apoJ-mediated suppression of immune complex-mediated glomerulonephritis. 158 11. Postulated regulatory role for apoJ in Th1 versus Th2
immune response differentiation. 160
4
General Introduction
5 ApoJ biological roles. The apolipoprotein J (apoJ) gene encodes an intriguing protein
with a striking expression pattern and numerous biological properties. ApoJ protein was first
purified from ram testes fluid, and characterized as a pro-cell aggregation factor (1). Since that
initial study, varied functions or processes have been attributed to apoJ activity based on its
tissue distribution and protein-binding partners. Proposed apoJ roles include complement regulation, cell death modulation, lipid transport, membrane protection, and protein chaperone
activity (2-7).
ApoJ expression appears to be a fundamental feature of many tissues undergoing
development or remodeling, suggesting that apoJ is an essential morphogenic factor. In vitro,
apoJ promotes cell aggregation of multiple cell types including spermatozoa, Sertoli cells,
erythrocytes, and renal epithelial cells (1, 8, 9). In vivo, apoJ may facilitate cell-cell and cell- substratum interactions during tissue development, and may maintain such contacts during tissue damage such as in renal tubular injury. In renal cell culture, cytoskeletal alterations that disrupt cell-cell and cell-matrix interactions are associated with increased apoJ message levels (9).
Sertoli cells induce apoJ expression at the time of cell attachment, and inhibition of apoJ activity correlates with anchorage-dependent apoptosis (10). ApoJ also promotes the formation of vascular smooth muscle cell nodules in tissue culture, which have been implicated in the pathogenesis of atherosclerosis (11).
Most bodily fluids are rich in apoJ protein, often coupled to various lipoprotein particles such as HDL and VHDL. The relationships between apoJ and lipoprotein particles suggest that apoJ is involved in lipid transport. Compatible with its developmental and injury-inducible expression pattern, apoJ may facilitate lipid recycling at dynamic tissue sites. ApoJ induces cholesterol and phospholipid efflux from macrophage-foam cells in vitro, and associates with the
6 secreted cholesterol in the culture medium (12). Thus, apoJ may suppress lipid-mediated diseases such as atherosclerosis by promoting lipid export.
ApoJ may modulate other lipoprotein activities besides lipid transport. HDL particles generally protect against tissue injury by suppressing inflammation and oxidative stress via the activity of various co-factors such as platelet-activating factor acetylhydrolase, ceruloplasmin, and paraoxonase (13, 14). ApoJ is induced in a wide array of inflammatory lesions and, in the context of HDL, may represent a supplementary anti-inflammatory co-factor. ApoJ levels are often inversely correlated with levels of the anti-oxidant HDL co-factor, paraoxonase, in response to oxidized lipids (15, 16). ApoJ is also co-deposited with paraoxonase in inflammatory atherosclerotic lesions (14).
Considering the hydrophobic binding nature of apoJ, the protein may clear lipid-based toxic molecules generated during developmental or pathologic apoptosis. Fibroblasts in cell culture induce apoJ expression upon exposure to a wide variety of cellular lipid debris, including apoptotic vesicles, disrupted cells, and membrane remnants (17). The hydrophobic bias of apoJ may also aid in the protection of cellular membranes against harmful bioactive fluids and complement-mediated cell lysis. This theory is compatible with the robust expression of apoJ along fluid-tissue boundaries and at pathologic interfaces between healthy and injured tissues
(18).
ApoJ has long been speculated to be a regulating factor of complement activity. Human apoJ protein was initially described in the context of glomerular immune complexes containing the soluble form of the MAC (SC5b-9) (19). In follow-up studies, it has been demonstrated that apoJ is an invariable component of such glomerular deposits. ApoJ also co-localizes with the
MAC in a variety of other lesions including myocardial and renal infarcts (20, 21). These
7 observations inspired a series of experiments investigating a possible role for apoJ in
complement regulation. ApoJ purified from serum binds several MAC proteins in vitro, and
inhibits complement-mediated cell lysis (2, 22, 23). In addition, serum depletion of apoJ
enhances complement-mediated injury in an isolated perfused kidney injury model (24).
Anti-complement activity is one of several postulated functions for apoJ in the male
reproductive tract, as serum and seminal fluid forms of apoJ have equivalent complement-
inhibiting properties (25). Seminal plasma and spermatozoa contain abundant quantities of
several apoJ protein isoforms, which appear to differ by site of synthesis, subunit composition,
and carbohydrate content (26-28). All seminal fluid isoforms of apoJ inhibit complement-
mediated lysis; however, apoJ from rete testis fluid is more effective than apoJ from cauda
epididymal fluid (26). The general complement regulatory activity of seminal fluid may be
essential for reproductive efficiency in response to allogeneic interactions in the female genital tract (29). Consistent with this theory, apoJ seminal levels and apoJ-positive spermatozoa directly correlate with increased fertility rates (25, 30). Besides complement regulation in the male reproductive tract, apoJ may sequester abnormal spermatozoa, and may transport lipids from developing spermatozoa (25, 31).
As has been the case with other proteins, it would appear that the diverse protein-protein interactions of apoJ in vitro are somewhat misleading. Rather than demonstrating multiple specific apoJ activities, the protein binding profile of apoJ suggests that apoJ has protein chaperone properties. The apoJ protein secondary structure confers non-specific protein binding with numerous molecules (32). In addition, one of the apoJ protein domains shares sequence homology with the chaperone domain of the small heat shock protein, αB-crystallin. And
comparable to many heat shock proteins, apoJ protein has a higher in vitro binding affinity for
8 denatured proteins than native proteins (3). This increased binding affinity prevents protein
aggregation during heat shock (33). Thus, apoJ is analogous to heat shock proteins based on its
gene regulation and protein function, and may act as a novel extracellular chaperone (32).
The ability to modulate protein aggregation has implications for apoJ in assorted diseases such as Alzheimer’s Disease (AD). ApoJ protein co-localizes with amyloid beta (Aβ) deposits
in AD brains, and interacts with a soluble form of Aβ (sAβ) in the cerebral spinal fluid (CSF)
(34). ApoJ also augments the transport of sAβ complexes across the blood-brain and blood-CSF
barriers, possibly via an interaction with the gp330/megalin receptor (35). In vitro, apoJ inhibits
Aβ aggregation and polymerization, and prevents Aβ-dependent cytotoxicity (36-38). These
observations all suggest that apoJ may protect against Aβ-mediated neurodegeneration. In line
with this hypothesis, apoJ levels in the hippocampus and neurofibrillary tangles, are inversely
correlated with severity of dementia (39, 40).
ApoJ expression is a hallmark of cell stress and tissue injury. Early studies examining
apoJ induction during tissue involution suggested that apoJ was a pro-apoptotic factor (6).
However, more recent investigations have demonstrated that apoJ is essential for cell survival.
Inhibition of apoJ activity in cell culture augments apoptosis in response to heat shock, oxidative
stress, and UV radiation in cell culture (5). The systemic administration of apoJ anti-sense
oligonucleotides induces tumor apoptosis and regression in the Shionogi mouse prostate tumor
model (41). On the contrary, prostate cancer cells that over-express apoJ protein are less
sensitive to cytotoxic agents. Thus, apoJ promotes cell survival, and may be exploited to
facilitate tumorigenesis and cancer progression.
The protective properties of apoJ are also represented at the tissue level. As previously
mentioned, low levels of apoJ in the CNS are associated with advanced neurodegeneration.
9 ApoJ serum levels are also inversely correlated with renal disease activity in lupus patients (42).
In the absence of apoJ, mice are more sensitive to autoimmune myocarditis (43). ApoJ-deficient
animals develop more severe myocardial lesions, and are more apt to present with heart failure,
than wild-type animals. These observations suggest that apoJ suppresses inflammatory and
immune-mediated tissue injuries.
ApoJ gene structure. ApoJ gene orthologs have been identified in all vertebrates
examined thus far. The gene of interest in our laboratory, the mouse apoJ gene, is located on
mouse chromosome 14, and maps to the mouse mammary tumor virus locus Mtv-11 (44). The
human apoJ homolog is located on human chromosome 8p21, adjacent to the lipoprotein lipase
gene (45). The mouse and human apoJ genes encompass approximately 20 kilobases (kb), and
contain 9 exons and 8 introns (Figure 1). The exonic structure of the mouse apoJ gene is entirely conserved with ortholog genes in quail, rat, and human species at the level of exonic organization and at the division of protein domains. In addition, there is an 81% nucleotide
sequence similarity between the mouse and human apoJ DNA-derived protein coding sequences
(44). The transcription start site precedes exon 1, and the polyA signaling motif is positioned in
the beginning of exon 9. The protein translation start site is located in exon 2, and the translation stop codon occurs at the end of exon 8 (7).
In the human apoJ gene, there are as many as seven polymorphisms, five of which are contained within the protein-coding sequence (46, 47). The apoJ gene polymorphisms are more common in the African/African-American populations than the Caucasian population. The polymorphisms are not known to be associated with any particular disease (47). However, the
polymorphisms may have an effect on different subpopulations of serum HDL (46).
10 My thesis advisor, Dr. Bruce Aronow, carried out a computer sequence analysis of apoJ
ortholog genes to identify conserved cis-elements within intronic and flanking gene regions. As previously observed by other investigators (48), the gene comparison demonstrated that the apoJ core promoter, encompassing 200 basepairs (bp) preceding the transcription start site, is highly conserved among the quail, mouse, rat, and human apoJ genes (Bruce Aronow, personal communication). Amid the conserved core promoter sequences are two key regulatory motifs, a heat shock element (HSE) and an activator protein-1 (AP-1) binding site.
The sequence analysis also revealed that regions of the distal mouse apoJ promoter, as
well as the mouse apoJ first intron, have high homology to corresponding human apoJ gene
regions (Bruce Aronow, personal communication). Additionally, the region of the mouse apoJ
exon 9 corresponding to the 3’-untranslated portion of the mRNA includes a series of sequence
motifs conserved in the equivalent region of the human apoJ gene (Bruce Aronow, personal
communication). Dr. Aronow also noted that human apoJ gene introns 1 and 6 have undergone
significant nucleotide expansions compared to the mouse gene.
ApoJ protein structure and protein-protein interactions. The mouse apoJ protein is
best characterized as a secreted glycoprotein consisting of two 40-kDa subunits (α and β) (49).
The subunits are linked together by five interchain disulfide bonds. The α and β subunits are
enzymatically cleaved from a common precursor protein, which is translated from a 2.0 kb apoJ
mRNA transcript. The α and β subunits are derived from the amino and carboxyl terminal halves
of the precursor protein, respectively. The mouse apoJ precursor protein cleavage site is a
conserved arginine-serine (Arg-Ser) bond, which is encoded within the 3’ end of exon 5. The
11 final protein product is composed of 449 amino acids, and has a molecular weight of 75-80 kDa
(50).
Based on structure and sequence analysis, the apoJ protein is comprised of multiple
functional domains (Figure 2). Secreted apoJ is initially synthesized with a consensus
hydrophobic signal peptide (2). The mouse apoJ signal peptide is 21 amino acids long, and is cleaved from the apoJ preprotein upon proteolysis in the endoplasmic reticulum (ER). The mouse apoJ protein also contains six potential N-linked glycosylation sites (7). The attached carbohydrates make up as much as 30% of the total mass of apoJ, and are often branched, complex, and abundant in sialic acid (3).
ApoJ contains several potential protein-protein binding domains, including four linear heparin-binding domains, two coiled-coil helices, and three amphipathic α-helices (32). ApoJ interacts with a diverse assortment of molecules including bacterial products, lipids, heparin,
Ku70, glutathione S-transferase (GST), apolipoprotein A-I, amyloid β peptide, glycoprotein (gp)
330/megalin, paraoxonase, immunoglobulins, and complement proteins C7, C8, and C9. The majority of these interactions are probably hydrophobic in nature, which correlates with the presence of the amphipathic α-helices in the apoJ protein (3). Amphipathic α-helices are proposed to stabilize both hydrophilic and hydrophobic interactions. Thus, the α-helical secondary structures are thought to be key domains with respect to the general protein binding capability and biological function of apoJ.
Though apoJ is most often considered a secretory protein, nuclear forms of apoJ have
been detected. Nuclear apoJ is induced in human liver HepG2 and mink lung CCL64 epithelial
cell lines in response to transforming growth factor-beta (TGF-β) treatment (51). The nuclear
apoJ protein lacks the signal peptide, and is thought to be generated via an in-frame secondary
12 start codon 33 amino acids downstream of the first start codon. Nuclear targeting is attributed to a SV40-like nuclear localization signal (NLS) in the apoJ protein. In MCF-7 human breast cancer cells, apoJ also localizes to the nucleus in response to ionizing radiation, where it interacts with the DNA double-strand break repair protein complex, Ku70/Ku80 (52).
ApoJ distribution and tissue expression. Mouse apoJ protein in adult animals is found in most bodily fluids, including semen, urine, CSF, breast milk, and serum (50). ApoJ protein levels in the seminal fluid are at a concentration of 500 to 1000 µg/ml. Western and immunofluorescence analyses demonstrate that multiple apoJ protein isoforms exist in seminal fluid, based on molecular weight and carbohydrate content (26-28). A native 80-kDa apoJ protein isoform associates with a minority fraction of spermatozoa that are characterized by immature or abnormal morphology (28). In contrast, a novel protein isoform, recognized only by anti-alpha chain antibodies, is localized within the acrosomal cap of normal spermatozoa.
In serum, the concentration of apoJ is 50 to 100 µg/ml, and the protein is associated with a subset of high and very high-density lipoprotein particles (HDL and VHDL) (4, 50). ApoJ-
HDL particles contain all key types of lipids, but have a reduced lipid content compared to the majority of HDL fractions (7). ApoJ-HDL particles also include cholesteryl ester transfer protein activity, paraoxonase, and apoA-I (4). ApoA-I binds apoJ, and incorporates apoJ into
HDL particles (53, 54). HepG2 human hepatocellular carcinoma cells synthesize an alternative form of apoJ-lipoproteins that lack apoA-I, and contain high levels of lipids, especially triglycerides (31). Rabbit retinal Muller cells and murine BV2 glial cells also secrete lipid-rich apoJ-lipoproteins (55, 56).
13 The majority of adult mouse apoJ-expressing cells share several characteristics (18).
First, most of the cell types are epithelial and highly secretory. Second, consistent with its
abundance in bodily fluids, apoJ-expressing cells often form boundaries between the underlying
tissue and biologically active fluids such as gastric acid, bile, and urine. In most organ systems,
apoJ expression is limited to a specific subset of epithelial cells. For example, in the
gastrointestinal (GI) system, only epithelial cells lining proximal portions of the GI tract
synthesize apoJ, such as in the gastric glands, Brunner’s glands, pancreatic ducts and glands, and
biliary ducts (18). GI expression of apoJ is not observed beyond the anterior duodenum.
In the male reproductive tract, apoJ synthesis is robust, but restricted to Sertoli cells along the seminiferous tubules, tubular epithelial cells in the head of the epididymis, and the seminal vesicles (27, 57). By comparison, apoJ expression is strong throughout the entire length of the female genital system, from follicular-lining granulosa cells in the ovary to vaginal mucosal cells (18). ApoJ production in the urinary system is more limited, and occurs in distal convoluted tubular epithelial cells in the renal medulla, and in urothelial cells lining the renal pelvis, ureters, and bladder (18, 58).
Constitutive expression of apoJ is not strictly limited to epithelial cells. For example in the heart, apoJ is synthesized in cardiomyocytes of the left and right atria, but not by ventricular myocytes (18). Central nervous system (CNS) expression of apoJ occurs in epithelial and epdendymal cells lining the choroid plexus and ventricles, respectively. Additionally, glial cells and neuronal subpopulations, such as the cerebellar Purkinje cells, transcribe apoJ. In the eye, the retina, the ciliary body and the Harderian gland generate large amounts of apoJ (18, 59).
ApoJ expression in the hematopoietic system is limited to megakaryocytes, which synthesize apoJ in abundance. Megakaryocytic apoJ is packaged in platelet α-granules, and is released
14 upon platelet activation (60). As a result, platelets are important source of apoJ at sites of tissue
injury and inflammation.
Besides constitutive expression, apoJ transcription appears to be a fundamental feature
during tissue remodeling. ApoJ levels are significantly elevated in the apoptotic epithelia of the
involuting ventral prostrate following hormone ablation (6, 61-63). In the mammary gland,
ductal and alveolar epithelial cells express apoJ during pregnancy, but repress synthesis during
lactation (64, 65). Upon mammary gland involution, apoJ is dramatically induced in epithelial
cells undergoing apoptosis. During blastocyst implantation, apoJ synthesis is increased in
uterine glandular and luminal epithelial cells (66). ApoJ message is also observed in uterine
stromal cells and myocytes during decidualization.
In the developing mouse embryo, apoJ mRNA is first detected at day 9 of gestation, and
becomes present in a variety of epithelial cell types, derived from all three germ layers (67, 68).
ApoJ expression is correlated with branching morphogenesis in the developing lungs and
kidneys, and is observed in most bronchioles and tubules, respectively (67). With further
development and differentiation, apoJ synthesis is completely repressed in the lung, and
restricted to distal tubules in the kidney medulla. In contrast to the adult heart, apoJ is synthesized in both the embryonic atria and ventricles (67, 68). A finer examination demonstrated that apoJ expression in the developing heart is localized to the truncus arteriosus and endocardial cushions.
ApoJ expression during cell stress and disease. In contrast to its restricted constitutive expression, numerous cell types synthesize apoJ in response to environmental insults. In cell culture, many forms of cell stress induce apoJ expression, including serum withdrawal, heat
15 shock, oxidative stress, UV and ionizing radiation, and hypoxia (5, 48, 52, 69). In the majority of these cases, apoJ induction is associated with later stages of the stress response, and the onset of apoptosis. This expression pattern suggests that apoJ modulates cell survival and/or apoptosis. In fact, inhibition of apoJ expression via anti-sense therapy augments apoptosis in response to heat shock and UV radiation (5). Thus, apoJ appears to promote survival in cell culture.
In vivo, apoJ induction and protein deposition occur in numerous disease processes ranging from neurodegeneration to cancer. ApoJ expression most often occurs at the interface between healthy and injured tissue. In the response to deafferenting and neurotoxic brain injuries, apoJ is induced in astrocytes and certain neuronal populations (70, 71). For example, with kainic acid treatment, apoJ transcription is increased in susceptible pyramidal neurons and adjacent astrocytes of the hippocampus (70). ApoJ levels are also elevated in vulnerable brain tissue following hypoxia-ischemia (72-76). Stroke-dependent induction of apoJ occurs in the peri-infarct area, 3 to 7 days following the initial insult.
Neurodegenerative conditions, including scrapie, sporadic amyotrophic lateral sclerosis,
Pick’s Disease, and AD are associated with increases in apoJ message (37, 77-79). AD has been a major apoJ research focus, as apoJ mRNA and protein levels are elevated in the frontal cortex, hippocampus, and entorhinal cortex in response to AD (37, 40, 80). ApoJ expression is observed both in neurons and astrocytes. In addition, the CSF concentration of apoJ protein is increased in
AD patients compared to control patients. Comparable to apoE, apoJ is found in most types of senile plaques of the cerebral cortex (81). However, apoJ is not present in apoE-containing diffuse deposits in non-demented elderly patients or young Down’s syndrome (DS) patients.
16 Atherosclerotic vessel plaques also contain abundant apoJ protein deposits (44, 60, 82).
Compared to appropriate control groups, plasma levels of apoJ are significantly increased in susceptible animals fed atherogenic diets, and human patients with defined coronary artery disease (15, 44). Absent in healthy vessels, apoJ is present in the intimal and medial layers of atherosclerotic vessels (82). Platelets, via degranulation, may be a major source of apoJ in the early stages of pathogenesis (60). Additionally, protective HDL lipid particles may deposit apoJ in the atheromatous lesions along with other molecules, such as paraoxonase and apoA-I (14).
Local foam cells may also synthesize apoJ (44).
ApoJ is induced in cardiac ventricular tissue following myocardial infarction (MI). In a rat MI model, apoJ expression occurs for up to two days following the ischemic insult (83).
ApoJ synthesis does not occur in infiltrating inflammatory cells, but rather is localized to myocytes and endothelial cells in the peri-infarct zone. A human clinical study demonstrated that apoJ protein co-deposits with the terminal complement complex (MAC) on injured cardiomyocytes in 8 to 14 day old lesions (20). Cardiac ventricles also synthesize apoJ in response to autoimmune and viral myocarditis (43, 84). ApoJ expression levels are directly proportional to disease severity, as apoJ mRNA and protein are detected in ventricular myocytes at the boundary between injured and healthy cardiac tissue. As with myocardial infarction, inflammatory cells do not induce apoJ expression.
A variety of renal conditions involve apoJ induction and deposition. Nephrotoxic injury models increase renal tubular expression of apoJ in correlation with injury severity (85-87).
Renal tubule epithelial cells also induce apoJ expression 24 to 72 hours after an ischemic insult
(88). Comparable to a myocardial infarction, apoJ protein is co-deposited with the MAC on tubular epithelia along the infarct lesion boundary (21). Additionally, tubular induction of apoJ
17 is a hallmark of all clinical and experimental renal cystic disorders (58, 89-92). ApoJ expression is localized to epithelial cells lining cysts, and is repressed upon experimental reversal of cystic disease. Independent of renal tubular expression, numerous glomerular diseases, including
membranous glomerulonephritis, systemic lupus erythematous, and IgA nephropathy, are
associated with apoJ-containing immune complex deposits (93, 94).
The links between apoJ expression and various cancers have recently gained scientific
attention. ApoJ induction has been observed in a diverse array of developing solid tumors
including prostate, ovarian, breast, kidney, and brain (95-101). ApoJ is also a tumor specific
marker for anaplastic large cell lymphomas (98). Intracellular apoJ mRNA and protein levels
often directly correlate with the progression of a benign to malignant tumor, and a primary to a
metastatic tumor (101). In contrast, apoJ levels inversely correlate with the extent of apoptosis
in tumors. In rodent models of prostate cancer, over-expression of apoJ increases tumor cell
resistance to chemotherapeutic agents (41). Conversely, apoJ-deficiency enhances tumor cell
susceptibility to cytotoxic agents. ApoJ appears to be an ubiquitous stress response factor for
both normal and abnormal cells.
ApoJ regulation. ApoJ is regulated primarily at the level of transcription. Most studies
have focused on the activity of the core promoter and adjacent 5’ flanking sequences (Figure 3).
The core promoter contains several conserved regulatory motifs, including a HSE at position –
120 (relative to the transcription start site) in the mouse gene, which is identical in the quail, rat,
and human apoJ genes (48). The HSE is an essential stress responsive cis-element for many
genes, and its presence in the apoJ promoter suggests that apoJ is a functional member of the
heat shock protein family. In vitro, the apoJ HSE is capable of binding the transcription factor,
18 heat shock factor 1, and by itself, the regulatory motif confers reporter gene induction following heat shock. In addition, mutation of the HSE in the context of the quail promoter abolishes heat shock dependent reporter gene activity.
Mammalian apoJ gene promoters also contain a conserved AP-1 motif, at position –77 in the mouse gene. AP-1 activity is strongly associated with numerous cellular stress responses, and may promote the diverse expression profile of apoJ. The regulatory motif interacts with several AP-1 related proteins in vitro, including c-fos, Fra1, Fra2, JunB, and JunD (102). The apoJ AP-1 cis-element confers reporter gene induction in response to TGF-β, epidermal growth factor (EGF), and nerve growth factor (NGF) in multiple cell types (102-104). The AP-1 binding element also mediates the constitutive expression of apoJ reporter gene in some experimental systems (104).
5’ flanking apoJ sequences include additional cis-elements for B-myb, nuclear factor
(NF1), and CBF transcription factors. Over-expression of the B-myb transcription factor promotes tumorigenesis and induces endogenous apoJ transcription (105). B-myb binds a regulatory motif at position –292 in the human apoJ gene, which confers B-myb reporter gene trans-activation. The NF1 protein family regulates multiple genes during mammary gland differentiation, and a 74-kDa NF1 isoform binds the mouse apoJ gene at two positions between –
317 and –353 (106). The interaction of the NF1 factor with the mouse apoJ gene is concurrent with gene induction during mammary gland involution. The transcription factor CBF interacts with an adjacent region spanning –436 to -302 in the rat apoJ gene that confers Sertoli cell- specific reporter gene expression (107). The Sertoli cell enhancer is further defined by a core element that contains Sp1 and Ets binding sites.
19 Based on comparison with the quail apoJ gene, the first intron of the mammalian apoJ
gene may also contain regulatory elements. The quail gene contains two promoters that both
facilitate gene transcription in a similar manner in vivo (108). Additionally, both quail promoters
individually drive heterologous gene expression. The first quail apoJ promoter is identical to the
characterized mammalian apoJ promoter. The second quail apoJ promoter contains several
candidate-binding sites that are conserved in the mammalian apoJ first intron, including an AP-1
motif, two TGF-β inhibitory binding sites, and a glucocorticoid receptor element.
Conclusions. The apoJ gene product is a conserved and ubiquitous factor among vertebrate organisms. The apoJ protein structure is optimized to facilitate interactions with a multitude of endogenous and foreign molecules. The apoJ gene regulatory circuitry is primed to induce gene expression in response to cell stress, as apoJ is synthesized and deposited at pathological interfaces in many disease processes. Additionally, apoJ deficiency augments apoptosis and tissue injury. Thus, with numerous properties of a heat shock protein, it is likely
that apoJ prevents cell death and limits tissue damage via its chaperone activity. Considering the
current ambiguity of its role in pathophysiology, the further study of apoJ gene regulation and protein function will yield many clues concerning how our bodies respond to and resolve disease
processes.
20 FIGURES
Figure 1. Map of the mouse apoJ gene and protein coding sequences. The top schematic represents the exonic-intronic structure of the mouse apoJ gene (drawn to approximate scale).
The bottom drawing demonstrates the corresponding DNA-derived protein sequences. The mouse apoJ gene encompasses approximately 20 kb, and contains 9 exons and 8 introns. The translation start codon is positioned in the second exon, and the translation stop codon is located at the end of the eighth exon. The apoJ protein signal peptide is encoded by the second exon.
The apoJ protein α and β subunits are encoded by the distal end of exon 2 through exon 5, and the distal end of exon 5 through exon 8 respectively. This drawing is copied from the work of
Jordan-Starck et al. (7).
21
mouse apoJ gene
1 2 3 4 5 6 7 8 9
protein coding sequences
signal α subunit β subunit peptide
Figure 1. Map of the mouse apoJ gene and protein coding sequences.
22 Figure 2. Postulated structure of the mammalian apoJ protein. The diagram depicts the predicted structure of the mature apoJ protein (drawn to approximate scale). The apoJ protein consists of 449 amino acids, and has a molecular weight of 75-80 kDa. The apoJ protein is composed of two 40-kDa subunits (α and β), which are connected by multiple disulfide bonds.
Computer sequence analysis suggests that the apoJ protein contains several protein-protein interaction domains including three amphipathic α-helices, two coiled-coil helices, and four heparin-binding domains. The apoJ protein also includes multiple sites for N-linked glycosylation; carbohydrates make up as much as 30% of the total mass of the apoJ protein, and differentiate tissue-specific apoJ isoforms. This drawing is adapted from the work of Jordan-
Starck et al., and Wilson and Easterbrook-Smith (7, 32).
23 α-subunit
β-subunit
amphipathic α-helix
coiled-coil helix
heparin binding domain
cysteine rich region
disulfide bond
glycosylation site
Figure 2. Postulated structure of the mammalian apoJ protein.
24 Figure 3. Regulatory factors of the mammalian apoJ gene promoter. The drawing demonstrates known transcription factor binding sites of the apoJ gene promoter region
(nucleotide positions –500 to 0, drawn to approximate scale) and their relevant signaling pathways or cellular processes. The core promoter contains a conserved heat shock element at position -120 that confers stress responsive gene expression. In addition, a conserved AP-1 motif at position –77 in the core promoter mediates growth factor dependent apoJ gene induction. Beyond the core promoter is a B-myb binding site at position –292 that modulates apoJ gene expression during tumorigenesis. At positions –317 and –352, NF1 binds the apoJ promoter and induces gene expression in response to mammary gland involution. And finally,
CBF interacts with a Sertoli cell specific enhancer located between positions –302 and –436.
25
Sertoli cell factor cell stress
tumorigenesis
CBF NF1 NF1 Bmyb HSF AP1 -500 TATA
tissue involution growth factors
Figure 3. Regulatory factors of the mammalian apoJ gene promoter.
26
Chapter 1
Cooperation between Multiple Cis-Elements of
the Mouse Apolipoprotein J Locus
Is Required for Inducible Gene Expression
27 ABSTRACT
Under basal conditions, the mammalian apolipoprotein J/clusterin (apoJ) gene is
synthesized in a restricted series of secretory epithelial cells that form fluid-tissue boundaries. In
tissue injury and disease processes, apoJ gene transcription is strongly induced. I sought to characterize the mechanisms responsible for the expression profile by examining the regulatory functions of the mouse apoJ promoter and downstream gene sequences. I assessed the activities of heterologous reporter genes containing various apoJ gene fragments in transfected and transgenic cells. I observed that an apoJ core promoter activator protein-1 (AP-1) binding site
was essential for constitutive and inducible reporter gene expression. The AP-1 motif also acted
in synergy with an apoJ core promoter heat shock element (HSE) to confer stress responsive
gene expression. Additionally, I verified that heat shock factor 1 (HSF1) directly activated apoJ gene transcription via the HSE. Reporter gene examination of downstream apoJ gene sequences revealed that intron 1 augmented gene activation. I also observed that distal apoJ cis-elements silenced basal reporter gene expression, and consequently increased the extent of induction.
Further dissection of the silencer region implicated DNase I hypersensitive sequences within intron 6. I concluded from my study that the mouse apoJ gene is regulated by multiple elements located throughout the gene sequence, including a well-conserved core promoter, a first intron enhancer, and a sixth intron silencer.
28 ACKNOWLEDGMENTS
Guang Zhu synthesized several of the reporter genes utilized in this study. She also
assisted in maintaining and genotyping the transgenic mouse colony. The University of
Cincinnati DNA Core sequenced the reporter gene vectors. The University of Cincinnati
Transgenic Mouse Core generated the transgenic mice. Hector Wong generously provided the
HSF1 deficient and wild-type fibroblasts, and the HSF1 expression vector.
29 INTRODUCTION
The eukaryotic genome is rigorously controlled such that only a fraction of genes get transcribed (109). The proper regulation of gene expression is essential for a diverse array of biological processes, ranging from developmental programs to cellular stress responses, and involves a complex interplay between numerous signaling pathways. The robust injury induction of the apolipoprotein J/clusterin (apoJ) gene suggests that it is an ideal candidate with which to study gene regulation in response to environmental and physiological stress resulting in apoptosis, tissue damage, and inflammation.
The apoJ gene encodes a heterodimeric glycoprotein that is postulated to have an extracellular chaperone function during tissue injury and remodeling (3, 32, 33, 110, 111).
Constitutive expression of apoJ in adult animals is mostly limited to secretory epithelial cells lining fluid-tissue boundaries (18). However, apoJ is markedly induced in apoptotic cells during tissue development and involution (63, 65, 66, 68, 112-115). ApoJ gene induction and protein deposition also characterizes numerous types of pathologic apoptosis associated with neurodegeneration, atherosclerosis, and immune complex-mediated disease (19, 37, 39, 44, 60,
76, 89, 93, 94, 110, 116).
ApoJ gene expression is intimately linked with the cellular stress response. The stress response includes a highly conserved network of factors involved in the detection and reaction to environmental changes (117, 118). The stress response is often defined as the heat shock response, which is a misnomer, as the stress response program reacts to a diverse array of environmental cues, including developmental processes, growth factor stimulation, oxidant injury, infection, inflammation, and tissue aging (118). These processes all predispose
30 intracellular protein misfolding and aggregation for which the stress response initiates heat shock
protein (HSP(s)) synthesis. The HSPs, via chaperone and protease functions, repair damaged proteins, and thus act to salvage cells from environmental stress.
ApoJ is hypothesized to be a HSP based on its biochemical and expression properties.
The apoJ protein contains three amphipathic helices and two coiled-coil helices, which mediate protein-protein interactions in a rather promiscuous fashion (33, 111, 119-121). One of the apoJ coiled-coil domains shares sequence homology with the chaperone domain of the small HSP,
αB-crystallin (32). Consistent with a chaperone function, apoJ protein has a higher in vitro binding affinity for denatured proteins than non-stressed proteins. This increased binding affinity prevents protein aggregation during heat shock (3, 33, 111). These biochemical qualities suggest that apoJ protein may have chaperone activity in vivo comparable to the HSPs.
ApoJ gene induction often parallels HSP gene expression during heat shock, ischemia/reperfusion, and irradiation in a variety of cell culture and whole body experiments (5,
48, 52, 72-74, 76, 83). In some cases, apoJ gene activation is delayed, but sustained during cell injury (69, 75). In contrast, HSP gene induction is almost always immediate and transient in
comparison. Hence, apoJ appears to function later in the stress response. Regardless of the
temporal kinetics, apoJ gene expression appears to contribute to cell survival, because anti-sense
therapy against apoJ enhances cell death in response to heat and oxidative stress (5). Anti-sense
apoJ oligonucleotides also induce tumor cell apoptosis in the murine Shionogi tumor model (41).
Regulatory examinations of apoJ gene expression have been restricted to the promoter
region and adjacent 5’ flanking sequences (48, 103, 104, 106, 107). The apoJ gene core
promoter contains several cis-elements that are linked to the stress response. One such DNA
binding motif is a heat shock element (HSE), which is identical over 14 basepairs (bp) in the
31 quail, rat, mouse, and human apoJ gene promoters (48). The mere presence of a HSE in the core
promoter can be used to define apoJ as a stress response gene, since the motif is found in almost
all eukaryotic HSP genes (32).
HSE motifs confer HSP gene induction via interactions with the heat shock transcription
factors (HSF(s)). The principal mammalian HSF is HSF1, which is homologous to yeast and
Drosophila HSF (117, 118). HSF1 activity is required to generate the heat shock response, and without HSF1 cells are very sensitive to environmental changes (122, 123). HSF1 synthesis is
constitutive and ubiquitous. In unstressed cells the protein exists in the cytoplasm as an inactive
monomer. With cell stress, HSF1 is activated to form a homotrimer, which translocates to the
nucleus, and binds HSE motifs to activate gene transcription.
The apoJ HSE alone is capable of mediating reporter gene induction in response to heat
shock (48). In addition, mutation of the quail promoter HSE abolishes reporter gene expression
in response to heat shock. The apoJ HSE is composed of three (nGAAn) pentanucleotide units
or motifs, in alternating orientations. This sequence structure is equivalent to the HSF trimeric
binding sites of other HSP genes (124, 125). Accordingly, the apoJ HSE can bind HSF1 protein
derived from heat shock cellular extracts in vitro (48). However, it has yet to be shown that
HSF1 directly activates apoJ gene transcription in vitro or in vivo.
The apoJ core promoter also contains a DNA binding motif for the activator protein 1
(AP-1). The AP-1 cis-element, like the HSE, is a prototypical transcriptional target during cell
injury, and is an essential motif for the expression of a multitude of stress related mammalian
genes (126-130). The AP-1 transcription factor is a homo or heterodimer composed of members
of the c-jun and c-fos protein families. Stress-activated signaling cascades, such as the Jun N-
terminal kinase (JNK) pathway, regulate AP-1 activity (131). These pathways are activated by
32 numerous environmental changes, including hypoxia, reperfusion, hyperthermia, and UV
irradiation (131, 132). Concerning heat shock, AP-1 activity alone is capable of inducing gene expression in some cases (129).
The apoJ AP-1 element is located immediately downstream of the HSE in the proximal promoter. The site is identical in the rat, mouse, and human apoJ homologues, but unlike the
HSE, the AP-1 element is not present in the quail promoter (48). The apoJ AP-1 element has been postulated to have many regulatory functions pertaining to cell injury, and growth factor/
hormone activity. Mutations of the rat and mouse apoJ AP-1 elements inhibit reporter gene
induction mediated by transforming growth factor beta (TGF-β), nerve growth factor (NGF), and
epidermal growth factor (EGF) in cell culture (103, 104).
The apoJ AP-1 motif binds multiple AP-1 protein components in vitro, including c-fos,
Fra1, Fra2, JunB, and JunD (102). This diverse protein binding profile confers a greater degree
of regulation via the AP-1 site. For example, c-fos represses basal apoJ gene expression in CCL-
64 mink lung epithelial cells. However, upon TGF-β exposure, c-fos no longer binds the apoJ
AP-1 element, and apoJ transcription is subsequently induced.
ApoJ gene regulation remains ambiguous beyond the few studies of its proximal promoter. This uncertainty exists in spite of its remarkable expression pattern in a wide array of tissue injury and disease states. We believe that this diverse profile can only be mediated by multiple factors acting in synergy to fulfill the actions of numerous signal transduction pathways.
To identify supplementary regulatory cis-elements, Cathy Ebert examined the chromatin structure of the entire mouse apoJ locus via DNase I hypersensitivity analysis. Her study revealed numerous accessible sites throughout the gene including the core promoter, intron 1,
33 and distal gene sequences (personal communication). Of particular interest were two tissue-
specific DNase I accessible sites within the first half of intron 6.
Utilizing these clues, Guang Zhu characterized a series of mouse apoJ reporter genes in
transfection experiments and transgenic animals. ApoJ 5’ flanking sequences, including the
proximal promoter, were sufficient to promote reporter gene expression in cell culture, but not in
transgenic animals (personal communication). However, upon the addition of intron 1, reporter
transgene expression was robust in many tissues. When apoJ-reporter transgenes were extended to include apoJ gene sequences past exon 5, reporter enzyme activity was ubiquitously repressed.
From these observations we concluded that the mouse apoJ gene is regulated by an intron 1 enhancer element, and a downstream silencer element.
In the present study I tested the hypothesis that the apoJ promoter HSE and AP-1 motifs work together to activate apoJ gene expression. I also sought to confirm the role of HSF1 on apoJ gene induction. I synthesized apoJ reporter gene constructs containing various promoter mutations, and then examined their responsiveness to growth factor treatment and heat shock in transfected and transgenic cells. I observed that the apoJ AP-1 cis-element is essential for both constitutive and inducible reporter gene expression. By comparison, apoJ HSE function was more selective, as heat shock inducibility, but not constitutive expression, was suppressed by the hse mutations. My study also established that HSF1 activated apoJ transcription via co- expression studies and gene expression analyses in HSF1-deficient cells. In addition, I observed that the ap-1 and hse mutations synergistically abolished heat shock-induced and HSF1- dependent reporter gene expression. Thus, I concluded that both the HSE and AP-1 cis-elements are required for the apoJ gene stress response.
34 I also evaluated the regulatory significance of other apoJ gene cis-elements identified by our laboratory in the above model systems. We hypothesized that the first intron enhancer augments gene activation, and that the downstream silencer acts like an on/off switch during the stress response, yielding the “all or nothing” expression profile of the endogenous gene. I observed that intron 1 did enhance TGF-β and heat shock-mediated reporter gene induction. The apoJ downstream silencer did not prevent reporter gene activation, nor did it alter the temporal kinetics of the reaction. Instead, the silencer increased the extent of induction by limiting basal gene transcription. Consequently, the observed changes in reporter enzyme activities were greatest in the presence of the silencer element. Further dissection of the apoJ silencer region implicated a DNase I accessible region in intron six, but the study failed to identify a specific motif(s) or transcription factor (s). In conclusion, my work suggests that the stress response induction of the mouse apoJ gene is generated by multiple, cooperating cis-elements located in the core promoter, first intron, and distal apoJ gene sequences.
35 MATERIALS AND METHODS
Construction of CAT expression vectors. DNA fragments of the mouse apoJ gene
were removed from the genomic clones lambda dash-63 and lambda dash-270 (44) and inserted
into pBLCAT6 to create ApoJ-CAT constructions. pBLCAT6 was the parent CAT expression
plasmid and contained the chloramphenicol acetyl transferase gene (CAT), and SV40 large T-
antigen intron and polyadenylation sequences (133).
The wild-type promoter and first intron-containing construction, CAT-1 WT, was
synthesized in two steps. First, a 1.4-kilobase (kb) Xba1-Sac1 fragment from intron 1 was
ligated into ApoJCAT-Ex2 between the Xba1 and Bgl2 sites. Additional intron 1, exon 1, and 5'
flanking sequences were added into the Xba1 site. A 3.8-kb Xba1 fragment was removed from a
BamH1 4.3-kb/PBSKS+ subclone. As a result of cloning, a 45-bp segment between the Sac1
and Nco1 sites in intron 1 was omitted. CAT-1 WT was made by Guang Zhu.
Mutated promoter and first intron-containing constructions (CAT-1 ap-1, CAT-1 hse1,
CAT-1 ap-1/hse1, CAT-1 hse2, CAT-1 ap-1/hse2) were derived from the CAT-1 WT and CAT-1
hse1 vectors. Site-directed mutagenesis of mouse apoJ gene proximal promoter cis-elements,
including an AP-1 element at mouse apoJ gene (NCBI accession # AF182509) positions 5476-
5482 and a HSE element at mouse apoJ gene positions 5433-5447, was carried out utilizing the
polymerase chain reaction (PCR). Separate fragments incorporating the mutations in either their
5’ or 3’ respective ends were isolated using PCR, ligated, and the subsequent fusion fragment was again amplified. CAT-1 wild-type sequences were then swapped for the mutated fusion
sequences via AvrII and AgeI restriction enzyme sites. Double stranded DNA sequencing
36 reactions, carried out by the University of Cincinnati DNA Sequencing Core, confirmed the identities of the mutated CAT-1 sequences. Anand Patak made the CAT-1 hse1 construct.
The HSF1 cDNA expression vector, HSF1, contained full-length human HSF1 cDNA
under the control of a constitutively active promoter. The construct was provided by the
laboratory of H. Wong (Children’s Hospital Research Foundation, Cincinnati, Ohio, USA), and
was originally generated by the laboratory of C. Wu (134).
The 3’ extended construction of CAT-1, CAT-2 WT, contained a 7.45-kb BsaH1-EcoRI
insert encompassing part of intron 2 through most of intron 7, which was placed 3’ of the
reporter gene polyA signal. Guang Zhu made CAT-2 WT.
The ADA promoter construction, ADA-CAT, was created from the 5’I23 ADA promoter
construct, and contained the human adenosine deaminase (ADA) minimal promoter (14).
The intron 6 SV40/ApoJ promoter-luciferase constructions (F(n)R(n)-SV40/ApoJ-Luc)
were synthesized using oligonucleotide primers corresponding to intron 6 sequences. The
generated series of fragments were subcloned into the Kpn1 site of the pGL3-luciferase vector
(Promega Corporation, Madison, WI, USA), upstream of the SV40 or apoJ minimal promoters.
The apoJ minimal promoter corresponded to a 2.1 kb BamH1 fragment derived from ApoJCAT-
4. Oligonucleotides (sequence corresponds to the mouse apoJ gene, NCBI accession #
AF182509): forward primer 5' ends: F1=13355, F2=13721, F3=14255, F4=14726, F5=14883;
reverse primer 3' ends: R1=14276, R2=14716, R4=16289, R5=14904.
CAT-2∆1 was derived from CAT-2 WT and contained a 1.194 kb deletion of the first
half of intron 6 corresponding to mouse apoJ gene (NCBI accession # AF182509) sequences
13717-14909. DNA fragments flanking the deletion were isolated by the polymerase chain
reaction (PCR) and cloned into the pPCR-Script vector separately (Stratagene, La Jolla,
37 California, USA). The fragments were then ligated together within the pPCR-Script vector
utilizing SalI and EcoRV restriction enzyme sites. CAT-2 sequences were removed upon AflII
and PvuI restriction enzyme digestion, and replaced with the new fusion fragment containing the partial intron 6 deletion.
Cell lines. Mouse liver Hepa 1-6 cells (ATCC CRL-1830) and NIH 3T3 cells (ATCC
CCL-92) were grown in Dulbecco's modified Eagle's medium (DMEM) (Gibco BRL, Rockville,
MD, USA) that contained 1% penicillin-streptomycin, L-glutamine, 110 mg/L sodium pyruvate, and 10% fetal bovine serum (FBS) (HyClone Laboratories, Inc., Logan, UT, USA) or calf serum
(CS) (Gibco BRL) respectively. Mink lung epithelial cells (Mv 1 Lu (NBL-7), ATCC CCL-64) were grown in DMEM/F12 medium (Gibco BRL) that included 1% penicillin-streptomycin, 1%
L-glutamine, and 10% FBS.
HSF1 wild-type and deficient fibroblasts were provided by the laboratory of H. Wong,
and were originally generated by the laboratory of I. Benjamin (122). The cells were grown in
high glucose DMEM medium that contained 0.1 mM MEM non-essential amino acids solution
(MEM NEAA) (Gibco BRL), penicillin-streptomycin, L-glutamine, 0.055 mM 2-
mercaptoethanol (BME) (Gibco-BRL), and 10% FBS.
Transgenic analysis of the mouse apoJ-CAT reporter genes. Restriction fragments
from the various reporter constructions were isolated from agarose gels and prepared for
microinjection (135). Transgenic mice were made with the fertilized eggs of FVB/N strain at the
University of Cincinnati Transgenic Mouse Core by the procedures described (136). Genotype
determination for the CAT reporter genes was initially performed via the polymerase chain
38 reaction (PCR), and confirmed by Southern analysis. Transgene copy numbers were determined by comparison of Southern blots, using EcoRI-cut liver DNA. To estimate transgene copy number, a 0.8-kb XbaI fragment from 5' flanking DNA was used to recognize a 4.8-kb EcoR1 fragment in the endogenous apoJ gene for quantitative comparison to a 1.4-kb band recognized by a 0.66-kb HindIII-ScaI probe from CAT coding sequences. Quantitative analysis of the relative intensities of the bands was accomplished by use of the PhosphorImager and
ImageQuant software. Each transgenic line contained multiple unrearranged copies of the reporter gene constructions.
Transgenic fetal fibroblasts (MFFs) were derived from mouse apoJ-CAT reporter transgenic fetuses (day 14 to 17). Specimens from representative transgenic lines for each apoJ-
CAT reporter construct were chosen for examination. The fetuses were eviscerated, minced, trypsinized, and cultured in DMEM complete media containing 10% FBS, 1% penicillin- streptomycin, L-glutamine, and 110 mg/L sodium pyruvate. Genotyping of the cultures was carried out by PCR and Southern analysis. The MFFs completed two cell culture passages prior to experimentation.
Transient transfection of cultured cells. The mouse apoJ reporter constructs were transfected into CCL-64, Hepa 1-6, NIH-3T3, and HSF-1 wild-type/deficient fibroblasts cells
(5.0x104 cells/35 mm plate or 2.5x106 cells/100 mm plate), using 5 µl of Lipofectamine (Gibco
BRL) per 1 µg DNA. Eqimolar amounts of the mouse apoJ reporter plasmids were co- transfected with pCMV-βgal (provided by the laboratory of S.S. Potter, Children’s Hospital
Research Foundation, Cincinnati, Ohio, USA) or pHBLacZ (provided by the laboratory of P.
Stambrook, University of Cincinnati, Cincinnati, Ohio, USA) reporter genes and pBluescript
39 (Stratagene) as neutral DNA for a total of 1 µg DNA/35mm plate or 5 µg DNA/100mm plate. In addition, multiple preparations of the reporter plasmids were used in most cases. The cells were incubated with the transfection reaction in serum free media (Opti-Mem, GIBCO-BRL) for 18-
20 hours at 37°C, and then replaced with DMEM complete media. After an additional 24 hours, the cells were harvested by rinsing the plates with 0.25% trypsin, 0.05 M EDTA, and then the cells were lysed either by three cycles of freeze-thawing (for CAT assays), or in 1X Reporter
Lysis Buffer (Promega), followed by one cycle of freeze-thawing (for luciferase assays).
TGF-β treatment. CCL-64 or NIH-3T3 cells were exposed to 5 ng/mL of human recombinant TGF-β1 (R&D Systems, Inc., Minneapolis, MN, USA), diluted in 4 mM HCl, in complete DMEM/F12 media, for up to 24 hours prior to harvest.
Heat shock treatment. Cell culture plates containing the various cell lines were partly submerged in water and exposed to 42°C for a sustained one to 24 hours prior to harvest. In some cases, cells were heat shocked for a determined amount of time, and then allowed to recover at 37°C until harvest.
Northern analysis. Total RNA was isolated from control or treated tissues/cells using
TRI-REAGENT (Molecular Research Center, Inc., Cincinnati, Ohio, USA). Total RNA (20µg) was separated by electrophoresis through a denaturing 1% agarose/formaldehyde gel and then transferred onto a nylon membrane. The ethidium bromide stained gel was photographed prior to the membrane transfer, and the image of the 18s rRNA band was quantified using ImageQuant software (Molecular Dynamics, Amersham Pharmacia Biotech, Sunnyvale, CA, USA) to
40 measure loading consistency. Following the transfer, CCL-64 RNA blots were probed with a 5’ human apoJ cDNA EcoRI fragment, and NIH-3T3 RNA blots were probed with a full-length mouse apoJ cDNA EcoRI fragment. The hybridized blots were then analyzed using a
PhosphoImager device (Molecular Dynamics) and ImageQuant software.
TGF-β/heat shock transient transfection analysis of the mouse apoJ-CAT reporter genes. CCL-64, NIH-3T3, and HSF1-deficient/wild-type cells were transfected with the mouse apoJ-CAT constructs as described above. TGF-β and heat shock treatments were carried out as previously described, beginning 48 hours after the start of the transfection, and lasting for up to
24 hours preceding harvest.
CAT, Lac Z, Luciferase, and protein determinations. Cell CAT activities were measured via the FAST CAT (deoxy) assay kit (Molecular Probes, Inc., Eugene, Oregon, USA), and quantified using a PhosphoImager device and the ImageQuant software package. CAT activities from transgenic samples were normalized for transgene copy number. Transfected cell β- galactosidase activities were quantified using the o-nitrophenyl-β-D-galactopyranoside assay
(ONPG) (Sigma Chemical Corp., St. Louis, Missouri, USA) (137). Total protein concentrations from tissue and cell extracts were determined using the Bradford assay (BioRad Laboratories,
Hercules, CA, USA). Luciferase reporter activities were measured using the Luciferase assay system (Promega), along with an automated luminometer (Bioscan, Inc., Washington, D.C.,
USA).
41 Reverse Transcriptase-Polymerase Chain Reaction (RT-PCR) Assays. Total RNA
(2.5µg), purified from control or treated cells using TRI-REAGENT, was incubated with 0.5 µg
of OligodT12-18 (Gibco BRL) at 70°C for 10 minutes. The primed RNA was transcribed in
reverse at 42°C for 50 minutes in reactions that contained 0.01 M dithiothreitol (DTT), 0.5 mM
of each deoxynucleotide triphosphate (dNTP), 1X reaction buffer, and 200 u of Superscript II
reverse transcriptase (Gibco BRL). The reverse transcription reactions were inactivated at 70°C
for 15 minutes, and the remaining RNA species were degraded using 2 u of RNAse H (Gibco
BRL) at 37°C for 20 minutes.
Mouse apoJ and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) cDNA were co-
amplified in PCR reactions that included 2.0 µl of RT template, 0.2 mM of each dNTP, 1.5 mM
MgCl2, 0.1 µM of each apoJ cDNA specific primer, 0.05 µM of each GAPDH cDNA primer,
and 0.8 u of Taq DNA polymerase (Gibco BRL). The PCR reaction profile consisted of: 94°C, 3 minutes, 1 cycle; 94°C, 30 seconds, 55°C, 30 seconds, and 72°C, 1 minute, 30 cycles; 72°C, 10
minutes, 1 cycle. The PCR products were visualized on 2.0% agarose gels, and then quantified
using the ImageQuant software.
Identification and analysis of conserved regions in mouse and human apoJ genes.
The mouse and assembled human apoJ genes were analyzed for regions of homology, using the
Advanced PipMaker program (138), with chaining option on and other settings at default, after having applied the Repeat Masker to the mouse sequence (139). The PDF file generated by
PipMaker was further annotated by using Photoshop (Adobe Systems, CA). The analysis of transcription factor binding sites in regions of high homology (70% or greater) was carried out with the MATInspector program (140) from Genomatix (Munich, Germany), using the Transfac
42 Professional Database (141) version 4.2 from Biobase (Braunschweig, Germany). Core
similarity was set to 1.00 and Matrix similarity was set to be the “calculated optimal”.
Independent of the Genomatix software, a separate computer program was utilized to detect and illustrate the occurrence of cis-elements that were conserved between the mouse and human genes (Aronow, Goud, and Sherwood, in preparation).
43 RESULTS
Mutational analysis of the mouse apoJ gene promoter. Prior examinations of apoJ
gene expression have implicated the core promoter, and in particular conserved HSE and AP-1
cis-elements, as key mediators of apoJ gene expression (48, 103). We hypothesized that both elements cooperate to confer apoJ gene activation in response to numerous environmental insults. To characterize the regulatory function of each site and their interdependent functions, I synthesized a series of mouse apoJ-CAT reporter genes containing specific promoter mutations in the context of additional 5’ flanking sequences, exon 1, intron 1, and exon 2 (Figure 1). The
HSE and AP-1 cis-elements were altered via missense mutations within their respective core binding sites, which were derived from earlier studies (Table 1) (48, 103). Two different HSE mutants were created, hse1 and hse2, respectively, to examine the mutational effects on
alternating HSE pentanucleotide motifs (Table 1).
I assessed the transcriptional effects of the mutations, both alone and combined, on
reporter gene constructs transfected into mink lung epithelial cells (CCL-64) and mouse
fibroblasts (NIH-3T3). ApoJ expression has been partially characterized in CCL-64 cells, but
has yet to be detailed in NIH-3T3 cells (103). I chose these contrasting cell lines in order to
dissect potential tissue and cell-specific regulatory programs for apoJ gene expression.
Additionally, basal apoJ expression is repressed in both cell types, making them optimal to study
apoJ gene induction.
The wild-type mouse apoJ reporter gene, CAT-1 WT, generated significant reporter
enzyme levels in both cell types, in contrast with the parental vector, pBLCAT (Figure 2).
However, the ap-1 mutation suppressed basal reporter gene expression in the CCL-64 and NIH-
44 3T3 cells. The hse1 core mutation did not alter reporter enzyme activity. On the contrary, the hse2 flanking pentanucleotide mutation enhanced reporter gene expression in NIH-3T3 cells,
both in the absence and presence of the ap-1 mutation.
I examined reporter gene responsiveness in two different apoJ gene induction model systems: TGF-β and heat shock treatment. It has been previously shown that apoJ mRNA and protein levels are elevated in CCL-64 epithelial cells by exposure to the growth factor, TGF-β.
To confirm this and further optimize the model system, I determined the extent of CCL-64 apoJ
gene transcription by utilizing Northern blot analysis (Figure 3). I compared apoJ gene
activation in response to sustained heat shock at 42°C versus TGF-β treatment. As expected,
apoJ mRNA levels were elevated as much as 35-fold by TGF-β in a time-dependent fashion. In
contrast, heat shock did not significantly increase apoJ mRNA levels in CCL-64 cells.
To dissect the mechanisms of TGF-β-dependent apoJ gene activation, I evaluated wild-
type and mutant apoJ reporter enzyme activities in CCL-64 transfected cells in response to the
growth factor (Figure 4). The pBLCAT parent vector was unresponsive to TGF-β treatment. In
contrast, the wild-type apoJ sequences in CAT-1 WT mediated considerable reporter gene
induction. The ap-1 mutation blunted reporter gene activation, thereby confirming the growth
factor response significance of the binding site. Mutations of the HSE element had no negative
effects on reporter gene expression.
Heat shock is another well-characterized agent of apoJ gene induction in many cell types.
I chose to investigate the heat shock response of the apoJ gene in NIH-3T3 cells based on
preliminary experiments that I had carried out (data not shown). As with the TGF-β study, I
initially examined endogenous gene expression levels in the absence or presence of stress via
Northern blot analysis (Figure 5). I examined apoJ gene activity upon exposure to sustained heat
45 shock at 42°C, oxidative stress in the form of hydrogen peroxide (H2O2), and TGF-β. In contrast
to the CCL-64 experiment, apoJ gene transcription in NIH-3T3 cells was greatly increased
during sustained heat shock. Oxidative stress and growth factor treatment had little effect on
apoJ gene activity.
I assessed the heat shock inducibility of the wild-type and mutant apoJ reporter genes
following transfection into NIH-3T3 cells (Figure 6). CAT-1 WT reporter enzyme activity was
significantly elevated during sustained heat shock compared to the pBLCAT. The hse1 and hse2 mutations equally and fully abolished heat shock-dependent induction, verifying the role of the
HSE motif in the stress response activation of the apoJ gene. Interestingly, the ap-1 mutation also moderately inhibited apoJ reporter gene expression in response to heat shock.
I compared the wild-type and mutant apoJ reporter gene expression profiles in transfected
NIH-3T3 cells during a time course heat shock study (Figure 7). For this study and subsequent experiments, I chose to omit the hse2 mutant constructs and focused on the hse1 mutant constructs. Following 1 hour (hr) of heat shock, the reporter gene levels were comparable to control levels (control data not shown). However, after 3 hrs of heat shock, the wild-type reporter gene was significantly induced. In contrast, mutant reporter gene levels were not elevated. CAT-1 WT reporter levels increased in a linear fashion following 6 hrs of heat shock.
There was also a slight induction of CAT-1 ap1 and CAT-1 hse1, but CAT-1 ap1/hse1 remained repressed.
Requirement of HSF1 during mouse apoJ gene heat shock induction. My studies confirmed that the apoJ HSE is essential for apoJ gene stress induction in NIH-3T3 cells. I next sought to determine if HSF1 directly activates apoJ gene expression via the HSE by analyzing
46 the effects of HSF1 co-expression on transfected wild-type and mutant apoJ reporter gene
transcription in NIH 3T3 cells (Figure 8). To modulate HSF1 activity, I varied both the amount
of co-expressed protein and incubation temperature.
HSF1 ectopic expression greatly increased CAT-1 WT reporter enzyme activity
compared to backbone vector pBLCAT. HSF1 trans-activation was dose dependent, and was
augmented at 42°C, consistent with prior studies (142-144). The hse1 promoter mutation
repressed HSF1 trans-activation at low HSF1 cDNA concentrations (0 and 2.5 ng ). However, at
high cDNA concentrations (25 and 250 ng) HSF1 was able to partially activate CAT-1 hse1 reporter gene expression. The ap-1 mutation had an even greater negative effect on HSF1 trans- activation, at both low and high HSF1 cDNA concentrations. In addition, the ap-1 and hse1
mutations synergized to inhibit HSF1 activity on apoJ reporter gene expression.
To determine the HSF1 requirement for apoJ gene activation in an endogenous setting, I
evaluated transfected CAT-1 WT reporter gene expression in HSF1-deficient and wild-type
immortalized mouse fibroblasts (Figure 9). I utilized two different formats of heat shock: 1) a 3
hr sustained heat shock followed by a 3 hr rescue, and 2) a 6 hr sustained heat shock with no
rescue. Without transfection, HSF1-deficient fibroblasts grew faster than their wild-type counterparts (data not shown). HSF1-deficient cells were also much more sensitive to sustained
heat shock compared to the wild-type fibroblasts, as symbolized by a greater frequency of
apoptotic cells (data not shown). At 37°C, CAT-1 WT reporter enzyme levels were rather
comparable in the two cell types, with a possible slight elevation in the absence of HSF1.
However, upon cell stress, the apoJ reporter gene was induced only in the wild-type cells.
I verified the apoJ expression differences in the two cell types via RT-PCR analysis
(Figure 10). I compared apoJ cDNA levels with that of the housekeeping gene, GAPDH,
47 following 0, 1 and 6 hr heat shock treatments. Without stress, apoJ transcription was non-
existent in wild-type cells, while HSF1-deficient fibroblasts generated a small amount of apoJ transcript. This observation correlated with the transfection data in HSF1-deficient cells and the expression results of the CAT-1 hse2 construct in NIH-3T3 cells. Following 1 hr of sustained
heat shock, there was a significant difference in apoJ transcription between HSF1-deficient and
wild-type cells. However, the apoJ transcriptional difference was less apparent after 6 hours of
heat shock treatment.
Regulatory roles of downstream mouse apoJ sequences. The transgenic
characterization of the mouse apoJ gene revealed the existence of an intron 1 enhancer element,
and a constitutive silencer element located beyond exon 5 (Guang Zhu, personal
communication). I investigated the role of the intron 1 enhancer during apoJ gene induction by
comparing the transfected reporter gene expression of CAT-1 WT versus CAT-0 WT, a CAT-1
WT derivative lacking the entire first intron (Figure 1, Figure 11). The reporter genes were
evaluated in the CCL-64/ TGF-β and NIH-3T3/ heat shock model systems (Figure 11). In both
systems, CAT-0 WT and CAT-1 WT basal reporter enzyme levels were equivalent. However,
upon TGF-β or heat shock treatment, CAT-1 WT reporter activity was significantly greater than
that of CAT-0 WT.
The evidence of a downstream silencer element regulating apoJ transcription was
intriguing, especially considering the inducible nature of the gene. We proposed that the injury
inducibility of the mouse apoJ gene might be fully conferred by a repressor-switch mechanism.
To test this hypothesis, I examined the CAT-2 WT reporter gene in the NIH-3T3 heat shock
model system (Figure 1, Figure 12). CAT-2 WT was derived from CAT-1 WT, and included
48 downstream apoJ sequences through intron 7. I also included in the study an adenosine deaminase (ADA) reporter gene to confirm the heat shock specificity of the apoJ constructs.
Both apoJ reporter genes were induced by heat shock in a time-dependent manner, while
ADA reporter gene expression did not change with heat shock. Based on apoJ gene Northern analysis (Figure 5), the expression profile of CAT-2 WT appeared more analogous to the endogenous gene than that of CAT-1 WT (Figure 12). Like the mouse apoJ gene, CAT-2 WT activity was completely silenced under basal conditions, but upon heat shock, the reporter gene was upregulated by approximately 40-fold. In contrast, CAT-1 WT was only induced about 10- fold, because of its higher constitutive expression.
Transgenic analysis of mouse apoJ reporter genes in response to heat shock. To confirm the physiologic relevance of the heat shock transfection experiments, Guang Zhu and I generated transgenic mice containing various mouse apoJ-CAT reporter gene constructs.
Subsequently, I isolated fetal fibroblasts from representative transgenic mice, and subjected the fibroblasts to sustained heat shock in cell culture (Figure 13). CAT-1 WT reporter transgene levels were significantly elevated with respect to heat shock time. Transgenic CAT-2 WT was induced in a greater fashion than transgenic CAT-1 WT after 4 hrs of heat shock, but curiously, the activation of the CAT-2 WT transgene was decreased following 13 hrs of heat shock. This observation may have been the result of cell death.
The hse1 mutation completely prevented transgene activation in response to heat shock.
In addition, the ap1 mutation partially suppressed heat shock-dependent transgene activation.
The negative effects of the ap1 mutation were more apparent after the initial 4 hrs of heat shock, as the fold induction of the CAT-1 ap1 transgene was half that of transgenic CAT-1 WT.
49 However, following 13 hrs of heat shock, the CAT-1 WT and CAT-1 ap1 reporter transgenes
were activated to a similar degree. Overall, these results were very consistent with the
transfection results, and confirmed the biological significance of the various apoJ regulatory
elements.
Characterization of the mouse apoJ gene silencer region. The downstream mouse
apoJ gene silencer emerged from my induction studies as a novel and interesting regulatory
element. Prior DNase I hypersensitivity and transgenic analyses of mouse liver tissue indicated
that intron 6 may contain the silencer element (Cathy Ebert and Guang Zhu, personal communication). To further delineate the mouse apoJ gene silencer, I synthesized a series of apoJ-luciferase gene constructs that contained various intron 6 sequences, including the two prominent tissue-specific DNase I accessible sites (Figure 14). The sequences were cloned upstream of either the SV40 or the mouse apoJ gene promoter. The reporter genes were then assayed in transfected Hepa 1-6 mouse hepatoma cells based on the previous liver-specific chromatin studies. I also chose to examine reporter gene expression in NIH-3T3 cells in order to explore potential cell-specific silencing mechanisms.
In the context of the SV40 promoter, intron 6 fragments containing both hypersensitive sites moderately repressed reporter gene expression in Hepa 1-6 cells, but not in NIH-3T3 cells.
However, upon the separation of the two tissue specific DNase I hypersensitive sequences, the repression was abolished in the liver cell line. In conjunction with a more powerful promoter, the
mouse apoJ promoter, the first half of intron 6 still moderately silenced reporter gene expression
in Hepa 1-6 cells. Additionally, the equivalent intron 6 region significantly repressed reporter
50 gene activity in NIH-3T3 cells. ApoJ-promoter constructs containing the isolated DNase I accessible sites were not examined.
Further dissection of the silencer region using this heterologous experimental system failed to yield any other clues about the exact location or identity of the negative regulatory element. Therefore, I continued my characterization of the silencer region via the mouse apoJ-
CAT reporter gene system. I removed the first 1.1 kb of intron 6 from the CAT-2 WT reporter construct, deleting the two DNase I accessible sites in the process, to create the construct CAT-
2∆1 (Figure 15). The expression effects of this deletion were analyzed upon transfection into
Hepa 1-6 and NIH 3T3 cell lines. As observed in previous transfection and transgenic experiments, downstream sequences extending past exon 5 silenced reporter gene expression.
The deletion of the putative silencer region partially abolished the repression in both cell lines.
51 DISCUSSION
The HSE and HSF1 mediate the stress response induction of the mouse apoJ gene.
My study highlighted the regulatory significance of the apoJ core promoter, and two of its most
highly conserved cis-elements, the HSE and AP-1 motifs. A previous investigation of the apoJ
promoter had identified the HSE, and characterized its function during the heat shock response
(48). My experiments confirmed those earlier findings, by indicating that a wild-type HSE was required for apoJ gene induction in response to stress. In contrast, the HSE was not essential for
TGF-β-mediated gene activation. Nor was the cis-element necessary for promoting basal reporter gene expression in the various model systems. Hence, the HSE regulatory role was specific for stress-mediated gene expression.
The prior analysis of the apoJ HSE also demonstrated that HSF1, derived from heat shock cellular extracts, was capable of binding the promoter sequence (48). To determine the
trans-activation potential of HSF1 via the HSE, I examined the effects of HSF1 co-expression
and deficiency on apoJ gene expression. I observed that changes in HSF1 cellular levels
significantly altered apoJ reporter gene activity in a wild-type HSE-dependent fashion, and that
HSF1 was required for the stress response induction of the apoJ gene. These results established that HSF1 was indeed a direct trans-activator of apoJ gene expression in cell culture.
It should be noted that the hse1 mutation did not completely abolish HSF1-mediated apoJ
reporter gene activation at high HSF1 concentrations. This observation could be explained by at
least three different possibilities: 1) indirect (non-DNA) binding activity of HSF1, 2) alternate
HSF1 DNA binding site(s), and 3) cooperative binding of HSF1. The last mechanism,
cooperative binding, has been identified in Drosophila HSF (145).
52 The optimal binding site for the Drosophila HSF trimer, a mammalian HSF1 homolog, is composed of three (nGAAn) pentanucleotide motifs, in alternating orientations, much like that of the apoJ HSE. Each HSF subunit binds an individual pentanucleotide motif. Thus, the three
HSF monomers act in synergy to increase the total binding affinity. If one of the pentanucleotide motifs is mutated, the Drosophila HSF binding affinity for its trimeric binding is reduced, but not completely abolished. This phenomenon occurs because the other two subunits cooperate to maintain DNA binding despite the absence of the third protein-DNA interaction. The cooperative binding nature of Drosophila HSF is even more apparent at higher protein concentrations, as was the case in my study. Therefore, the hse mutant apoJ reporter constructs may still respond to HSF1 since two of the three potential DNA binding motifs remain intact.
The two different hse mutations engineered in my study generated similar detrimental effects on reporter gene heat shock induction. However, it is interesting that mutation of the first pentanucleotide motif augmented basal reporter gene expression in NIH-3T3 cells. This enhanced expression even occurred in the context of an ap-1 mutation. I also observed a slight, but reproducible apoJ mRNA elevation in HSF1-deficient fibroblasts compared to their wild- type counterparts. These results suggest that the HSE and/or HSF1 may silence basal apoJ gene transcription in a cell-type specific fashion.
In the absence of cellular injury, HSF is negatively regulated by multiple mechanisms, in part because the stress response antagonizes cell growth and differentiation (117, 118, 142).
Under non-stress conditions, HSF1 binds DNA in vitro, but the factor is not capable of activating gene expression (117). Only upon heat shock, is HSF fully active to enhance transcription.
Inactive HSF1 may bind the apoJ HSE and repress gene expression by inhibiting basal transcription activators and/or altering the chromatin architecture of the apoJ promoter. Its
53 inability to bind, as in the mutant promoter experiments, or its absence, as in the HSF1-deficient experiments, relieves repression allowing increased basal apoJ gene expression. Contradictory to this hypothesis, HSF1 co-expression did not repress apoJ reporter gene transcription at control temperatures. I speculate that the experimental conditions, including the toxic transfection procedure and the excessive protein synthesis may stress the cells, and thereby activate HSF1.
An alternate mechanism may involve another member of the mammalian HSF family, such as HSF2. In contrast to HSF1, HSF2 is not involved in the stress response; but rather,
HSF2 activates gene transcription during development and differentiation (117, 118). HSF1 and
HSF2 are quite divergent in function and structure, except for their DNA binding domains, which are highly conserved (124, 125). Therefore, HSF1 and HSF2 may compete with each for binding the apoJ HSE binding. And in the absence of HSF1, HSF2 would have full reign of the apoJ HSE. In addition, the expression discrepancy between the hse1 and hse2 mutant constructs under non-stress conditions may be explained by decreased HSF1 binding and/or increased
HSF2 binding.
Despite the presence of a fully functional HSE in its core promoter, apoJ gene induction in response to heat shock is often delayed, but unremitting, compared to the acute transcriptional burst of other HSP genes. In human epidermal A431 cells exposed to constant heat shock, hsp70 gene induction occurs within 1 hr (69). The hsp70 mRNA levels peak at 4 hrs, and are completely diminished by 12 hrs of extended heat. In contrast, apoJ gene induction is not detectable until after 4 hrs of constant heat shock and is sustained for at least the next 20 hours in response to the stress.
The retarded activation of the apoJ gene, in comparison to other HSP genes, may be related to the apoJ HSE structure. The apoJ HSE contains only three pentanucleotide consensus
54 motifs. In contrast, many other HSEs include a greater number of motifs, which predisposes the
binding of added HSF trimers (124, 145). The additional binding of HSF molecules has
implications for induction kinetics based on the cooperative nature of HSF protein.
HSF DNA binding is increased by interactions with other HSF trimers, such that HSF
binding is directly correlated to the number of consensus motifs. Accordingly, HSF is most apt
to interact with the HSE with the greatest number of motifs when HSF protein amounts are
limited. The apoJ HSE, with only three motifs, may be unable to compete with larger HSEs for
the primary HSF1 bolus during the acute stress response, and thus, apoJ gene induction is
initially delayed until the protein is equilibrated.
Additionally, as with the other studies, heat shock-induced apoJ reporter gene expression
was maintained and increased over time in my experiments. The persistent stress induction of the apoJ gene suggests that supplementary factors in addition to HSF1 may modulate apoJ gene
activation. This idea was exemplified by apoJ gene induction in HSF1-deficient fibroblasts, and
CAT-1 hse1 activation in NIH-3T3 cells, both following 6 hrs of heat shock. With extended
injury or recovery, HSF1 activity is repressed by HSPs such as HSP70 or HSP90 in the manner
of a negative feedback loop (117, 118). Therefore, based on HSF1 protein regulation, it is likely
that apoJ transcription is independent of HSF1 during the later stages of my heat shock
treatments.
The expression differences between apoJ and other HSP genes may also symbolize a
functional distinction. HSP proteins such as HSP70 are considered early stress response factors.
On the contrary, apoJ induction may represent a late stress response during a critical cell
transition from survival to death. In line with this theory, the stress-mediated induction of apoJ
in HSF1-deficient cells was directly associated with accelerated cell death compared to wild-type
55 cells (data not shown). Thus, during cell death, apoptotic-specific transcription factors may promote apoJ gene expression.
The AP-1 motif is essential for mouse apoJ gene expression. AP-1 is a compelling candidate for a complementary transcription factor involved in the stress-mediated induction of the apoJ gene induction. The apoJ AP-1 cis-element has been hypothesized to mediate much of the apoJ gene expression profile. However, previous studies have only demonstrated that the
AP-1 motif is necessary for growth factor-dependent apoJ gene induction (103, 104). Consistent with their conclusions, I verified the AP-1 motif requirement in my TGF-β experiments.
Prior investigations also examined the role of AP-1 activity on basal apoJ gene expression, and were inconclusive. Jin and Howe observed that an ap-1 mutation, in the context of 1.3 kb of the rat apoJ promoter, does not alter basal reporter gene expression in four different cell lines including CCL-64 cells (103). In contrast, Gutacker et al. demonstrated that disruption of the same element in a 300 bp fragment of the mouse apoJ promoter represses basal reporter gene expression in PC12 cells (104).
Utilizing the same mutagenesis strategy as in the Jin and Howe study (103), I observed that the apoJ AP-1 motif, in the context of 1.4 kb of mouse apoJ 5’ flanking sequence, intron 1, exon 1, and exon 2, was required for basal reporter gene expression in CCL-64 and NIH-3T3 cells. It is possible my results and the similar Gutacker study (104) differ from the Jin and Howe study (103) because of promoter species differences, mouse versus rat respectively. Such an explanation is dubious because the mouse and rat 5’ flanking sequences are highly similar (Bruce
Aronow, personal communication). The inclusion of additional apoJ sequences in my experiment may have somehow increased transcriptional dependency for the AP-1 motif. This is
56 doubtful, and in fact, the basal requirement for the AP-1 binding site in a larger reporter gene
context strengthens the biological significance of the motif.
Although AP-1 trans-activation is highly associated with environmental insults, the apoJ
AP-1 motif had not been previously tested during the stress response. Therefore, I wanted to
determine if the cis-element was required for apoJ heat shock induction. The apoJ AP-1 cis-
element was essential for both heat shock-dependent and HSF-1-mediated apoJ reporter gene
activation, as the ap-1 mutation inhibited stress-induced reporter gene expression simultaneously
with the hse mutation at the onset of gene induction. Hence, the HSE and AP-1 binding sites
cooperate to induce early apoJ gene transcription. The marked repressive effects of the double ap-1/hse1 mutation in the heat shock time course and the HSF1 co-expression experiments
further demonstrated the synergistic relationship of the two core promoter motifs. These
findings correlate with previous AP-1 examinations, which established the cooperative nature of
AP-1 with other stress-related factors (130, 146-148).
My results suggest that a wild-type AP-1 promoter motif is vital for all modes of apoJ
gene expression. Constitutive AP-1 activity may promote an accessible chromatin environment
for both basal and inducible apoJ gene transcription (Figure 16). AP-1 factors have the innate
ability to bind their consensus site when it is reconstituted into a nucleosome, which
consequently permits the binding of secondary transcription factors (149). AP-1 also recruits
chromatin-remodeling complexes such as SWI/SNF to gene promoters (150). These AP-1
activities may be crucial for the binding of subsequent apoJ trans-activators including HSF1,
thereby explaining their synergistic relationship. Additionally, the activating role of AP-1 may
be expanded and/or amplified in response to heat shock and growth factor treatment via the
stress-activated JNK pathway.
57 Intron 1 enhancer augments mouse apoJ gene induction. Earlier regulatory studies of
the mouse apoJ gene identified an intron 1 enhancer element that was required for reporter
transgene expression (Guang Zhu, personal communication). I sought to characterize a possible
role for this enhancer in my apoJ gene induction model systems. In the absence of stress, intron
1 had little effect on the basal expression of the transfected reporter genes. However, in response
to growth factor treatment and heat shock, intron 1 significantly enhanced reporter gene
induction mediated by the apoJ promoter. Thus, the enhancer may increase apoJ gene expression
in a general manner. These novel findings imply that the core promoter is capable of, but not
solely responsible for apoJ gene activation. Rather, the apoJ promoter may act in synergy with
the first intron enhancer to fully induce apoJ gene transcription.
Though the cis-element dissection of intron 1 has not been completely carried out, some
clues pertaining to its putative enhancer function can be derived from a study of the quail apoJ
gene (108). The quail gene contains two promoters, the first of which corresponds to the known
mammalian apoJ promoter, while the alternative promoter has identity with sequences of the
mammalian apoJ first intron. Both promoters appear to drive apoJ gene expression in multiple
avian tissues in an equivalent fashion. In addition, both quail promoters mediate heterologous
gene expression independent of each other in a non-synergistic manner. Computer sequence analysis of the quail alternative promoter and its corresponding region in the mammalian apoJ first intron has identified numerous candidate binding sites including an AP-1 motif, two TGF-β inhibitory binding sites, and a glucocorticoid receptor element (151). Thus, the quail alternative promoter may have evolved into an unprecedented mammalian intronic enhancer.
58 Mouse apoJ gene downstream silencer element confers a greater degree of
regulation. Mouse apoJ basal gene expression is limited primarily to a subset of secretory
epithelial cell types (18). However, in response to a wide array of stressful conditions and
disease processes, apoJ is induced in many tissues and cell types. Based on our previous
chromatin and transgenic examinations of the mouse apoJ gene, we hypothesized that a
downstream silencer element mediates this striking difference between the basal and the
inducible expression of the apoJ gene via a repressor-switch mechanism.
My experimental results suggest that a negative element within a DNase I hypersensitive
region of intron 6 may restrict, at least partially, apoJ gene activity. However, in contrast to our
working hypothesis, the downstream silencer element is not required for gene induction. Nor did
it affect the induction temporal profile. Yet, the silencer is essential in increasing the range of
inducibility of the apoJ reporter genes, as determined by the fold induction differences between the silenced apoJ reporter gene, CAT-2 WT, and the active reporter gene, CAT-1 WT (Figure
12). It should be noted that the repression conferred by the silencer is never completely
abolished during gene induction, suggesting that the repressing mechanism is still partially
active.
Transcriptional de-repression has often been ignored as a mechanism for gene activation in response to cell stress. However, there are increasing numbers of examples in which genes are induced by inactivating transcriptional repressors. One such model gene is the Saccharomyces
cerevisiae hsp26 heat shock gene. The hsp26 promoter contains both activating and silencing cis-
acting elements (152). The deletion of the silencer elements de-represses the constitutive
transcription of hsp26 reporter constructs, and abolishes the full effects of heat shock induction.
It is hypothesized that the silencer elements are required to regulate hsp26 expression in a
59 physiologic manner. A similar conclusion can be made from my study, as the apoJ silencer
element enhanced reporter gene induction in response to heat shock.
The negative regulatory element may have evolved to decrease the level of apoJ gene
expression in the absence of tissue injury, thus insuring low basal amounts of apoJ protein. For
example, elevated apoJ levels during non-stress conditions may be detrimental to normal cellular
processes, similar to HSPs (142). In addition, apoJ gene activation is directly correlated with malignant transformation of certain cancer types (98, 99, 101, 153). ApoJ may promote tumorigenesis, possibly by preventing tumor cell apoptosis or blocking complement-mediated tumor cell lysis. Therefore, a silencer element may restrict apoJ expression during early stages of neoplasia, and indirectly limit apoJ-dependent transformation.
The intron 6 negative regulatory element may be analogous a classical silencer based on the definition of Ogbourne and Antalis (154). Though intronic in origin, the ability of the apoJ silencer to repress heterologous reporter gene expression argues against an mRNA splicing regulatory mechanism. Rather, since the silencer functions in a position-independent fashion, it is quite likely that the element confers active transcriptional repression. The observation that the silencer is promoter-dependent in a cell-specific fashion is intriguing. This phenomenon suggests that there are cell-specific factors for the apoJ silencer that act via different
mechanisms.
The inability to further delineate the intron 6 silencer element may be related to either
the chromatin state or cell nature of my experimental systems. The immortalized cell lines
utilized in my experiments may not generate a fully repressive chromatin structure, since
immortalized cells are undergoing constant DNA replication and proliferation. The maximum transcriptional repression contributed by the silencer element may also require target genes in a
60 natural chromatin context in contrast to non-physiologic episomes (155). This requirement is true for other repressor systems such as the Groucho protein model in Drosophila. Groucho represses the transcription of multiple genes via a histone-dependent mechanism, and hence can only inhibit the expression of genes integrated into chromatin (156, 157).
Computer sequence analysis of intron 6 indicates the potential regulatory significance of the region. The silencer element has high sequence similarity between the mouse and human
apoJ sequences based on a percent identity plot (Figure 17A). Of particular interest are the
blocks of intron 6 sequence, adjacent to the DNase I accessible sites 9.2 and 9.6, which exceed
75 percent identity. These DNase I hypersensitive regions are rich in conserved transcription
factor binding motifs (Figure 17B). With this information, it is interesting to note that separation
of the two DNase I accessible sites inactivated the silencer. This observation suggests that the
two sites may work in a cooperative manner to silence expression, and/or form a repressor
complex spanning the elements.
One particular factor motif identified, C/EBPβ, is a strong candidate for an apoJ
regulatory factor based on its function and expression profile. By homo- and heterodimerization
with itself and other C/EBP protein family members, the factor can mediate both transcriptional
activation and repression, as influenced by isoform concentration and cis-element context (158).
C/EBPβ is abundant in liver, where it acts as a key regulatory factor of acute phase proteins, a
class of inflammatory proteins that includes apoJ (159). The transcription factor is also induced
in many other tissues during inflammation, and in certain development processes, such as tissue
involution (160). A C/EBP factor may mediate the constitutive repression conferred by the apoJ
silencer element. Additionally, C/EBPβ may also inactivate the apoJ silencer and subsequently
aid in gene induction.
61 Future studies. My study revealed multiple novel facets of apoJ gene regulation, and
subsequently has generated a new series of experimental questions. I did not resolve the factor
binding status of the apoJ AP-1 cis-element during basal and inducible gene expression. A
preliminary experiment demonstrated that c-fos co-expression enhances apoJ basal reporter gene
expression in NIH-3T3 cells (data not shown). An additional pilot study indicated that heat
shock-induced apoJ reporter gene expression is not altered in c-jun-deficient fibroblasts (data not
shown). These experiments need to repeated and expanded to investigate the effects of all AP-1
protein family members during the various stages of apoJ transcription.
It may be interesting to further dissect the synergistic relationship between AP-1 and
HSF1 in an in vitro transcription system utilizing purified factors and/or cellular extracts. To determine the order of events during apoJ gene induction, one could add in a specific step-wise
fashion the general transcription factors and trans-activators onto an apoJ promoter DNA
template, and then measure protein-protein interactions and the transcriptional initiation rate.
Additionally, the template could be assembled with chromatin to determine the role of chromatin
in the activation mechanism, as well as to analyze the chromatin-remodeling effects of the trans-
activators.
The discrepancy of apoJ heat shock inducibility between the CCL-64 epithelial cells and
the NIH-3T3 fibroblasts implicates cell-specific mechanisms or factors, and could provide the
perfect experimental opportunity to further dissect the synergistic interactions of apoJ gene
effector proteins. I doubt that the HSF pathway is involved in this expression difference, considering that the heat shock response is a highly conserved and ubiquitous eukaryotic cellular program. Rather, I suspect that alternative transcription factors and/or associated signaling cascades are responsible for the induction difference between the two cell types.
62 It is intriguing that the ap-1 mutant reporter gene was moderately induced in mouse fibroblasts, but only following long-term heat shock. This was especially evident in the
transgenic cells. Theoretically, AP-1 is unable to bind this mutant construct, and HSF1 is
inactivated in later stages of the stress response. Thus, it is likely that an additional fibroblast
transcription factor activates apoJ reporter gene expression under these experiment conditions.
This observation emphasizes the need to fully mine the promoter, the intron 1 enhancer, and the
downstream silencer region for all possible DNA-binding motifs pertinent to apoJ gene
induction. Accordingly, a new series of reporter gene constructs will be required to analyze
novel cis-elements in cell culture and transgenic animals.
As my work was primarily based in cell culture, future endeavors should involve the
characterization of the various apoJ gene cis-elements utilizing relevant in vivo models of apoJ
gene induction, such as ischemia-repefusion or tissue involution (65, 74), for which HSF1 and
AP-1 activity have already been linked respectively (161-166). In addition, the apoJ mutant
transgene expression should be examined during development and organ differentiation. For
example, HSF2, via the HSE motif, may activate the developmental expression of the apoJ gene.
HSF2 synthesis and dependent gene activation occurs in several embyronic and adult tissues that
coincidentally express high levels of apoJ including the male reproductive tract, the central
nervous, and the heart (167-171).
Conclusion: Multiple cis-elements cooperate to regulate apoJ gene induction. Given
the significance of the numerous examples of distal cis-acting gene elements, I believe that the
regulatory circuitry of the mouse apoJ gene can only be understood by characterizing the entire
locus (135, 172-174). To think that a core promoter solely regulates a typical mammalian gene
63 is naïve. Recent data generated from the human genome project suggests that species evolution
is generated not by protein-coding gene diversion, but rather is promoted by alternative
regulatory sequences (175). Thus, it appears that the greatest genetic changes between species
occur in flanking and intronic sequences that contain putative regulatory elements.
My experimental observations of the mouse apoJ gene indicate that its transcription is
regulated by multiple cis-elements located throughout the gene locus (Figure 18). Previous
investigators have identified several promoter motifs that are essential for apoJ gene expression
(48, 103, 104, 106, 107). I extended those analyses by demonstrating that the two most
conserved apoJ promoter elements, the HSE and the AP-1 motifs, act in synergy to confer stress
responsive reporter gene expression. In addition, I confirmed that HSF1 is a direct trans-
activator of apoJ gene expression.
My study also established that an intron 1 enhancer augments apoJ reporter gene
induction, and that a downstream silencer confers a more inducible state by repressing
constitutive gene activity. Therefore, the basal and inducible expression profiles of apoJ may
result from a balance between gene activating mechanisms generated by the promoter and the
intron 1 enhancer, and gene repressing mechanisms mediated by the downstream silencer. By
fully examining and dissecting the regulatory mechanisms of the mouse apoJ gene, we will gain a much better understanding of multiple complex processes pertaining to tissue injury and cell death.
64 FIGURES AND TABLES
Figure 1. Schematic diagram of mouse apoJ gene sequences and reporter constructs. The mouse apoJ gene was dissected utilizing a series of reporter gene constructs, illustrated below the endogenous gene (drawn to approximate scale). The mouse apoJ gene proximal promoter, depicted above the full-length gene, contains a heat shock element (HSE) and an AP-1 binding site. The reporter gene, CAT-1 WT, contained 1.4 kb of wild-type 5’flanking mouse apoJ gene sequence, the first exon, the first intron, and the second exon with start codon, fused to the chloramphenicol acetyl transferase (CAT) reporter gene. The HSE and AP-1 elements are depicted by the black and white squares respectively. The CAT-1 ap-1 reporter gene was derived from CAT-1 WT and contained a 2 bp missense mutation of the AP-1 cis-element, as symbolized by the large “X”. The CAT-1 hse1 and CAT-1 hse2 reporter genes, also derived from
CAT-1 WT, included two different 3 bp missense mutations of the HSE element, as symbolized by the large “X”. The CAT-1 ap-1/hse1 and CAT-1 ap-1/hse2 reporter genes contained mutations of both the HSE and AP-1 elements, as symbolized by the two large “X”s. The reporter gene,
CAT-2 WT, was constructed using CAT-1 WT, and included additional downstream apoJ gene sequences through intron 7.
65
proximal promoter HSE AP-1 TATA
mouse apoJ gene 1 23456 789
CAT- 1 WT construct CAT
1 2
CAT- 1 ap-1 construct X CAT 1 2
CAT- 1 hse1 CAT CAT- 1 hse2 constructs X 1 2
CAT- 1 ap-1/hse1 CAT CAT- 1 ap-1/hse2 constructs XX 1 2
CAT- 0 WT construct CAT
1
CAT CAT- 2 WT construct 1 234567
Figure 1. Schematic diagram of mouse apoJ gene sequences and reporter
constructs.
66 Table 1. Wild-type and mutant mouse apoJ gene promoter element sequences. The mouse apoJ proximal promoter contains a heat shock element (HSE) and an AP-1 element, both of which differ by only 1 bp (depicted in lower case) from their respective consensus sequences
(depicted in UPPER CASE). To test the regulatory function of each site, 3 or 2 bp missense mutations of the cis-elements were generated in the context of the CAT-1 WT reporter gene
(missense mutations denoted in bold lowercase).
67
Table 1. Wild-type and mutant mouse apoJ gene promoter element sequences.
cis-element sequence
consensus heat shock element (HSE) NTTCNNGAANNTTCN mouse apoJ gene HSE CTTCCAGAAAGcTCC hse1 mutant element CTTCCAtccAGcTCC hse2 mutant element CaggCAGAAAGcTCC
consensus AP-1 element TGACTCA mouse apoJ gene AP-1 element TGAgTCA ap-1 mutant element TGActgA
68 Figure 2. Basal expression profiles of transfected mouse apoJ gene wild-type and mutant reporter constructs. Mouse apoJ reporter genes were transfected into CCL-64 and NIH-3T3 cells, and reporter enzyme levels were assessed in the absence of cellular stress. A constitutive
B-galactosidase reporter construct was co-transfected as an internal control for transfection efficiency. Reporter activity is defined by: CAT activity (percent conversion/ µg of protein extract/minute per reaction assay)/ β-galactosidase units, normalized to pBLCAT activity per cell line. The data points represent the mean reporter activities from three independent experiments.
The error bars symbolize the standard errors of the means.
69
construct
pBL CAT-1 CAT-1 CAT-1 CAT-1 CAT-1 CAT-1 CAT WT ap-1 hse1 ap-1/hse1 hse2 ap-1/hse2
14
CCL64 12 3T3
10
8
6 reporter activity
4
2
0 3t3
Figure 2. Basal expression profiles of transfected mouse apoJ gene wild-type and
mutant reporter constructs.
70 Figure 3. Northern blot analysis of apoJ gene induction in mink lung epithelial cells (CCL-
64). CCL-64 epithelial cells were subjected to sustained heat shock at 42°C or TGF-β exposure
(5 ng/mL 12 or 24 hours), and then examined for apoJ gene induction via Northern blot analysis
(right top image). The right bottom image depicts ethidium bromide stained 18s rRNA for the corresponding Northern blot samples. The left panel illustrates the quantified apoJ mRNA levels, as detected from the Northern blot, and corrected for loading inconsistency by using 18s rRNA levels. Experimental note: the apparent Northern hybridization signal in the TGF-β vehicle control lane is non-specific, and does not represent apoJ mRNA.
71
r r C 12 hr C 0 0 r r 42 37 TGFB 12 h vehicle control TGFB 24 h 40 r C C 12 h 0 0 TGFB 12 h 37 TGFB 24 h 42 vehicle control 30
apoJ
20
fold Induction 18s
10
0
Figure 3. Northern blot analysis of apoJ gene induction in mink lung epithelial cells (CCL-64).
72 Figure 4. TGF-β responsiveness of transfected wild-type and mutant mouse apoJ reporter
genes in CCL-64 cells. CCL-64 epithelial cells were transfected with the mouse apoJ reporter
genes, and then treated with either vehicle or TGF-β (5 ng/mL, 24 hours). A constitutive β-
galactosidase reporter construct was co-transfected as an internal control for transfection
efficiency. Reporter activity is defined by: CAT activity (percent conversion/µg of protein
extract/minute per reaction assay)/β-galactosidase units, normalized to basal pBLCAT activity.
The data points represent the mean reporter activities from three independent experiments. The error bars symbolize the standard errors of the means. The values cited below each reporter construct represent the ratio of reporter activity in the presence versus absence of TGF-β.
73
construct pBL CAT-1 CAT-1 CAT-1 CAT-1 CAT-1 CAT-1 CAT WT ap-1 hse1 ap-1/hse1 hse2 ap-1/hse2
TGFβ - + - + - + - +++- - - + 50
40
30
20 reporter activity
10
0
1.8 2.9 1.9 3.4 1.8 3.4 1.9 Ratio: (TGFβ /ctrl)
Figure 4. TGFβ responsiveness of transient transfected wild-type and mutant
mouse apoJ gene reporter constructs in CCL64 cells.
74 Figure 5. Northern blot analysis of apoJ gene induction in mouse fibroblasts (NIH-3T3).
NIH-3T3 fibroblasts were treated with sustained heat shock at 42°C, hydrogen peroxide (250
and 500 µM), or TGF-β (5 ng/mL, 8 hours), and then assessed for apoJ gene induction via
Northern blot analysis (right top image). The right bottom image depicts ethidium bromide
stained 18s rRNA for the corresponding Northern blot samples. The left panel illustrates the
quantified apoJ mRNA levels, as detected from the Northern blot, and corrected for loading
inconsistency by using 18s rRNA levels.
75
2 2 0 0 2 2 r
r C C 8 h 0 0 250 u M H 500 uM H 37 TGFB 8 h 42 vehicle control
2 12 2 0 0 2 2 r r C 8 h
10 C 0 0 500 uM H 42 vehicle control 37 250 uM H TGFB 8 h
8 apo J
6
18s fold induction 4
2
0
Figure 5. Northern blot analysis of apoJ gene induction in mouse fibroblasts (NIH-
3T3).
76 Figure 6. Heat shock inducibility of transfected wild-type and mutant mouse apoJ gene
reporter constructs in NIH-3T3 cells. NIH-3T3 fibroblasts were transfected with the mouse
apoJ reporter genes, and then subjected to heat shock (42°C, 12 hours). A constitutive B-
galactosidase reporter construct was co-transfected as an internal control for transfection
efficiency. Reporter activity is defined by: CAT activity (percent conversion/ µg of protein
extract/ minute per reaction assay)/ β-galactosidase units, normalized to basal pBLCAT activity.
The data points represent the mean reporter activities from three independent experiments. The error bars symbolize the standard errors of the means. The values cited below each reporter construct represent the ratio of reporter activity in the presence versus absence of heat shock.
77
construct
pBL CAT-1 CAT-1 CAT-1 CAT-1 CAT-1 CAT-1 1 ap-1/hse1 2 ap-1/hse2 CAT WT ap-1 hse hse
HS - + - + - + - + - + - + - + 180
160
140
120
100
80
reporter activity 60
40
20
0 1.8 27.0 10.0 1.8 1.8 1.7 2.0 Ratio: (HS /ctrl)
Figure 6. Heat shock inducibility of transient transfected wild-type and mutant
mouse apoJ gene reporter constructs in NIH-3T3 cells.
78 Figure 7. Time course plot of transfected wild-type and mutant mouse apoJ reporter gene expression during heat shock at 42°C. NIH-3T3 fibroblasts were transfected with the mouse apoJ reporter genes, and then subjected to heat shock at 42°C for 1, 3, or 6 hours. A constitutive
B-galactosidase reporter construct was co-transfected as an internal control for transfection efficiency. Reporter activity is defined by: CAT activity (percent conversion/µg of protein extract/minute per reaction assay)/ β-galactosidase units. The data points represent the mean reporter activities from two independent experiments. The error bars symbolize the standard errors of the means.
79 10
9 CAT-1 WT CAT-1 ap1 8 CAT-1 hse1 7 CAT-1 ap1/hse1
6
5
reporter activity 4
3
2
1
0 1234 56 sustained hours of heat shock
Figure 7. Time course study of transient transfected wild-type and mutant
mouse apoJ reporter gene expression during heat shock at 42°C.
80 Figure 8. HSF1 trans-activation of transfected wild-type and mutant mouse apoJ gene
reporter constructs in NIH-3T3 cells. Mouse apoJ reporter genes were co-transfected with
increasing amounts (0, 2.5, 25, 250 ng) of a HSF1 cDNA expression vector into NIH-3T3
fibroblasts. A constitutive β-galactosidase reporter construct was also transfected, as an internal
control for transfection efficiency. The transfected cells were then incubated at 37°C or 42°C for
6 hours. Reporter activity is defined by: CAT activity (percent conversion/µg of protein
extract/minute per reaction assay)/β-galactosidase units, normalized to basal pBLCAT activity.
The data points represent the mean reporter activities from three independent experiments. The error bars symbolize the standard errors of the means.
81 37°C 42°C 600 ng of co-transfected HSF1 cDNA 500 0 2.5 25 250
400
300 reporter activity
200
100
0 CAT-1 CAT-1 pBLCAT CAT-1 WT CAT-1 CAT-1 CAT-1 pBLCAT CAT-1 WT CAT-1 ap-1 ap-1/hse hse ap-1 ap-1/hse hse
1 1
construct 1 1
Figure 8. HSF1 trans-activation of transient transfected wild-type and
mutant mouse apoJ gene reporter constructs in NIH-3T3 cells.
82 Figure 9. Heat shock induction of the mouse apoJ CAT-1 WT reporter gene in HSF1- deficient and wild-type mouse fibroblasts. HSF1 deficient and wild-type fibroblasts were transfected with the CAT-1 WT reporter gene, and then incubated at 37°C, or 42°C for 3 hours followed by 37°C for 3 hours, or 42°C for 6 hours. Reporter activity is defined by: CAT activity
(percent conversion/µg of protein extract/minute per reaction assay)/β-galactosidase units, normalized to basal pBLCAT activity. The data points represent the mean reporter activities from two independent experiments. The error bars symbolize the ranges of reporter activities between the two experiments.
83 5 HSF -/-
HSF +/+ 4
3
fold Induction 2
1
0 37 °C 42 °C 37 °C 42 °C
Figure 9. Heat shock induction of the mouse apoJ gene CAT-1 WT reporter construct in HSF1-deficient and wild-type mouse fibroblasts.
84 Figure 10. RT-PCR analysis of mouse apoJ gene heat shock induction HSF1- deficient and wild-type mouse fibroblasts. HSF1-deficient and wild-type cells were subjected to sustained heat shock at 42°C for 1 or 6 hours, and then assessed for mouse apoJ gene induction via RT-
PCR analysis (top image). The bottom panel illustrates apoJ mRNA levels in the corresponding
PCR gel samples exactly above the respective data points, corrected for experimental variability by using GAPDH cDNA levels.
85
HSF-/- 42° C, 1 hr HSF+/+ 37° C (-) genomic DNA control control (-) water HSF -/- 37° C HSF -/- 37° (+) PCR control HSF+/+ 42° C, 1 hr HSF-/- 42° C, 6 hr (+) RT-PCR control 100 bp ladder HSF+/+ 42° C, 6 hr 100 bp ladder
apoJ
GAPDH 6
5
4
3
2 apoJ/GAPDH 1
0
Figure 10. RT-PCR analysis of mouse apoJ gene heat shock induction in HSF1-
deficient and wild-type mouse fibroblasts.
86 Figure 11. Role of mouse apoJ first intron enhancer on apoJ gene stress induction. CCL-
64 epithelial cells (A) or NIH-3T3 fibroblasts (B) were transfected with pBLCAT, CAT-0 WT,
or CAT-1 WT, and then treated with either TGF-β (5 ng/mL, 24 hours) (A), or sustained heat
shock (42°C, 12 hours). A constitutive β-galactosidase reporter construct was co-transfected as
an internal control for transfection efficiency. Reporter activity is defined by: CAT activity
(percent conversion/µg of protein extract/minute per reaction assay)/β-galactosidase units,
normalized to basal pBLCAT activity. The data points represent the mean reporter activities from three independent experiments. The error bars symbolize the standard errors of the means.
87
A construct pBLCAT CAT-0 WT CAT-1 WT
β TGF CCL- 50 - + - + - +
40
30
20 reporter activity
10
0 B construct pBLCAT CAT-0 WT CAT-1 WT
- + - + - + HS NIH- 180 3T3 160 140 120
100 80
reporter activity 60 40 20
0
Figure 11. Role of intron 1 enhancer during stress induction.
88 Figure 12. Role of mouse apoJ downstream silencer on stress-mediated induction. A) NIH-
3T3 fibroblasts were transfected with either CAT-1 WT or CAT-2 WT, and then subjected to multiple hours of heat shock at 42°C. A CAT reporter construct containing the human adenosine deaminase gene promoter was examined as a negative control for heat shock inducibility.
Reporter activity is defined by the: CAT activity (percent conversion/µg of protein extract/minute per reaction assay). Each data point is the mean of four assays (duplicate samples from two independent transfections). The error bars symbolize the standard deviations for each set of reactions. B) The diagram depicts the extent of reporter gene repression and heat shock induction conferred by the mouse apoJ promoter and first intron, in the presence or absence of the apoJ silencer contained in CAT-2 WT. The values are derived from figure A, at time points
0 (basal expression) and 24 hours (induction).
89 A 700 CAT-1 WT 600 CAT-2 WT ADA-CAT
500
400
300 reporter activity reporter 200
100
0 04812162024 sustained heat shock (hours) B
10 fold induction Basal apoJ reporter gene expression
42 fold induction
7 fold repression
Basal apoJ reporter gene expression with silencer
Figure 12. Role of downstream apoJ gene silencer on stress-mediated induction.
90 Figure 13. Transgenic analysis of mouse apoJ reporter gene expression in response to heat shock. Mouse fetal fibroblasts were isolated from representative transgenic mice, cultured for two passages, and then subjected to 0, 4, or 13 hours of sustained heat shock at 42°C. Fold induction is defined by the extent change of reporter gene levels following heat shock compared to control reporter gene levels, for each reporter construct respectively. The data points represent the average fold induction value for two independent reactions per construct per time point. The error bars symbolize the range of fold induction per construct per time point.
91 10 ° Hrs at 42 C 9 0 hrs
8 4 hrs 13 hrs 7
6
5 fold Induction 4
3
2
1
0 CAT-1 WT CAT-1 ap1 CAT-1 hse1 CAT-2 WT
construct
Figure 13. Transgenic analysis of mouse apoJ reporter genes in response to
heat shock.
92 Figure 14. Reporter gene dissection of mouse apoJ gene intron 6. A series of mouse apoJ reporter constructs were synthesized by inserting different segments of the putative silencer region into a reporter vector containing the luciferase gene and the SV40 (S) or mouse apoJ (AJ) promoters (drawn to approximate scale). The apoJ-luciferase reporter genes were transfected into either Hepa 1-6 or NIH-3T3 cells. A constitutive B-galactosidase reporter construct was co- transfected as an internal control for transfection efficiency. Reporter activity is defined by: relative light units (RLU)/β-galactosidase units. The data points represent the mean of at least four independent transfection reactions from one representative experiment. The error bars symbolize the standard deviation for each set of reactions.
93 DNase I hypersensitive sites 6 7 mouse apoJ gene
S luc NIH 3T3 Hepa 1-6
6 S luc
6 S luc
S luc
S luc
S luc
AJ luc
AJ luc
0 4000 8000 12000 reporter activity S = SV40 promoters AJ = ApoJ
Figure 14. Reporter gene dissection of mouse apoJ gene intron 6.
94 Figure 15. Characterization of mouse apoJ gene reporter gene with partial intron 6
deletion. The initial 1.1 kilobases (bp) of the mouse apoJ gene intron 6, including the liver
specific DNase I sites, were deleted from the CAT-2 WT reporter gene (depicted by the large
“X”) to create CAT-2 ∆1 (drawn to approximate scale). Reporter gene expression profiles were
then analyzed by way of transfection assays in either Hepa 1-6 or NIH-3T3 cells. A constitutive
β-galactosidase reporter construct was co-transfected as an internal control for transfection
efficiency. Reporter activity is defined by: CAT activity (percent conversion/µg of protein
extract/minute per reaction assay)/β-galactosidase units. The data points represent the mean of at least four independent transfection reactions from one representative experiment. The error bars symbolize the standard deviation for each set of reactions.
95 1 23456 789 mouse apoJ gene
CAT 1 2
CAT- 1
CAT 1 2 3 45 6 7
CAT- 2
NIH 3T3
Hepa 1-6
CAT 1 2 3456 7
CAT- 2∆1
0 200 400 600 800
reporter activity
Figure 15. Characterization of mouse apoJ reporter gene with partial intron 6 deletion.
96 Figure 16. Promoter-mediated induction of the mouse apoJ gene. (A) In the absence of basal AP-1 activity, nucleosomal arrays (white ovals, not drawn to scale) may block transcription factor-DNA interactions. (B) Upon AP-1 basal activation and its interaction with DNA, the chromatin structure of the mouse apoJ promoter may open, (C) facilitating the binding of additional factors such as HSF1 during gene induction. JNK may be a crucial player linking heat shock and growth factor treatment with amplified AP-1 activity (symbolized by larger arrows).
97 A
HSE AP1 TATA
B BASAL ACTIVATION
JUN FOS
HSE AP1 TATA
C HEAT SHOCK GROWTH FACTORS
JNK
HSF-1 JUN FOS
HSE AP1 TATA
Figure 16. Promoter-mediated induction of the mouse apoJ gene.
98 Figure 17. Computer sequence analysis of the mouse apoJ gene intron 6 silencer. The
percent identity between the human and mouse apoJ sixth introns, corresponding to mouse apoJ
gene silencer region, is plotted in panel (A) (drawn to approximate scale). The y-axis of the plot represents percent identity. The x-axis depicts the mouse sequence, 5’ to 3’, from left to right.
Each block (or line (as in the case of exon 6)) in the plot symbolizes a stretch of human
sequences that aligns with the mouse sequence, at a particular percent identity. Of potential
physiologic significance are the blocks of sequence with percent identities greater than 75%.
Exon 6 (rectangle) and the intron 6 tissue-specific DNase I hypersensitive sites (arrows) are
illustrated above the plot. B) Sequence analysis of the mouse versus the human apoJ gene,
specifically the DNase I site, 9.2, demonstrates that multiple cis-element motifs are aligned and
identical. The diagram in panel (B) depicts the mouse apoJ sequence (left vertical line), DNase I
site 9.2 (residues 14075-14275), aligned with the comparable sequence in the human gene (right
vertical line). The gene sequence is displayed from 5’ to 3’, top to bottom. The colored bands
connecting the two gene sequences symbolize identical protein binding motifs. The motifs were
identified using the Genomatrix database, containing previously characterized transcription
factor cis-elements. Only those sites with a core identity of 0.95 were included in the analysis.
99 A B HS 9.2 HS 9.6 EXON 6
100%
C/EBP MYO D
SRY SRY
75% FREA FREA
MYO D C/EBP
50%
Figure 17. Computer sequence analysis of the mouse apoJ gene intron 6 silencer.
100 Figure 18. Multiple cis-acting elements regulate the mouse apoJ gene. The diagram illustrates mouse apoJ gene cis-regulatory elements characterized in this and past studies (drawn to approximate scale). My experiments highlighted the cooperative interactions between HSF1 and AP-1 on the proximal promoter during the stress response. Additionally, the intron 1 enhancer element and a downstream silencer element, possibly within intron 6, increased gene inducibility. Considering the regulatory significance of the intronic enhancer and silencer sequences, I propose that apoJ gene expression is mediated by an equilibrium between gene activation and repression.
101
core general activator promoter region general silencer elements region
1 2 3456 789
Figure 18. Model of mouse apoJ gene regulation mediated by multiple putative cis-elements.
102
Chapter 2
Suppression of Immune Complex-Mediated Glomerulonephritis
by Mouse Apolipoprotein J
103 ABSTRACT
Apolipoprotein J (apoJ) is a secreted glycoprotein whose strong induction has been
observed in numerous pathologic situations. I tested the hypothesis that apoJ functions to
suppress tissue damage in immune complex-mediated glomerulonephritis using horse spleen
apoferritin immunizations in wild-type and apoJ-deficient mice. ApoJ-deficient mice
predominantly developed global proliferative glomerulonephritis, with an increased incidence of
hematuria-associated progressive disease, especially in females. Wild-type mice were
comparatively resistant to immune complex-mediated injury, presenting in most cases with either no or segmental glomerular lesions. Immunohistochemical and electron microscopic analyses demonstrated similar immune complex deposits with regards to localization and composition in both genotype backgrounds. However, serum antibody measurements indicated that apoJ- deficient animals developed a more polarized Th2 immune response than their wild-type counterparts. My results indicate that apoJ plays an anti-inflammatory role in response to immune complex mediated kidney injury, diminishing necrotizing inflammatory sequellae.
Given its ability to directly interact with both cells and hydrophobic proteins, I hypothesize that
apoJ inhibits inflammation by regulating immune complex interactions with key effector
molecules, thereby modulating host immune responses.
104 ACKNOWLEDGMENTS
Guang Zhu generated the apoJ-deficient mice. David Chmilelewski measured blood urea nitrogen and urinary creatinine levels. Lisa McMillin prepared the histologic specimens, and the glomerular pathology was assessed under the guidance of David Witte. David Finkel and Jean
Synder carried out the immunohistochemical and electron microscopic assays, respectively.
Judy Bean and Robert Tamer provided assistance with the statistical analyses.
105 INTRODUCTION
Despite advances in the understanding of renal disease pathogenesis, the incidence of renal disease has been rising over the past decade. The increased incidence of glomerular renal disease in the face of recent discoveries stresses the challenges in dissecting these diverse and complex injury processes. Glomerular disease is thought to be primarily immunologic in origin
(176). Glomerular immune complex (IC) deposits are postulated to confer tissue injury by at least six non-overlapping pathogenic pathways. The pathways include non-inflammatory humoral mechanisms and inflammatory mechanisms that involve a variety of cellular mediators
(177). Each pathway exhibits complex regulation affecting immune injury.
We propose that the protein, apolipoprotein J (apoJ), modulates both the cellular and humoral effectors of immune complex disease to protect tissues against further injury. ApoJ is a conserved, secreted heterodimeric protein with an intriguing pattern of expression and inducibility. The induction and/or deposition of the protein is associated with a variety of physiological and pathological processes including complement regulation, programmed cell death, neurodegeneration, atherosclerotic vascular disease, and renal injury (2, 6, 19, 44, 79).
The induction of apoJ during pathological conditions seems to occur in cells both within and surrounding lesions (75, 83). ApoJ is also expressed constitutively and developmentally in an interesting series of epithelial cells, frequently located at fluid-tissue interfaces (18). Thus, apoJ is a gene expressed at critical biological interfaces in both normal and pathophysiologic circumstances.
ApoJ synthesis in normal kidney tissue is restricted to cells within distal tubules (58).
However, apoJ expression is induced in numerous experimental and clinical renal diseases,
106 including cystic disease, hypo/dysplasia, and kidney transplant rejection (89). In the majority of
these cases, apoJ protein is localized to the tubular epithelium in a heterogeneous distribution.
ApoJ gene induction also occurs in renal tubules in a murine model of systemic lupus erythematous (SLE) (116). In an alternate IC glomerulonephritis (ICGN) rat model, glomerular transcription of apoJ increases compared to control animals, and the subsequent protein co- deposits with complement components (178).
Human apoJ protein was discovered initially in glomerular IC deposits isolated from a
patient with membranous glomerulonephritis (19). In follow-up studies apoJ was characterized
as a frequent component of glomerular deposits containing immunoglobulins and the soluble
form of the terminal complement complex (SC5b-9) in all types of renal pathology (93, 94). The
terminal complement complex, also known as the membrane attack complex (MAC), causes cell
lysis by forming transmembrane channels (176). The MAC is composed of the complement proteins C5b, C6, C7, C8, and C9, and its generation occurs either on a phospholipid membrane
(MC5b-9) or in fluid phase (SC5b-9) (19). MC5b-9 is cytolytically active via membrane attachment by complement C7, and subsequent polymerization of complement C9. In contrast,
SC5b-9 is cytolytically inactive, due to increased complex hydrophilicity and C9 polymer inhibition conferred by a supplementary protein, vitronectin. Despite its inactive state, SC5b-9 is a significant marker of various immunological diseases such as SLE (179, 180) .
The observation that apoJ is associated with SC5b-9 in human disease has inspired a series of investigations analyzing the role for apoJ in IC-mediated injury and complement regulation. ApoJ protein can bind IgA, IgM, all isotypes of IgG, and interact with both the Fc and Fab IgG domains in vitro (119). ApoJ also promotes IC aggregation under similar conditions (120). In addition, apoJ interacts with the C9 binding site of C5b-8 and the C5b-9
107 complex in vitro (23). ApoJ protein can be extracted from serum IgG and serum SC5b-9, suggesting that the aforementioned immunoglobulin and complement interactions do indeed occur in vivo (2, 119, 120, 181).
The experimental results pertaining to the complement regulatory role of apoJ have been mixed, partly due to in vitro assay variability. Early studies, utilizing purified MAC proteins, indicated that human apoJ inhibits complement-mediated erythrocyte lysis, but only prior to the addition of complement C7 and MAC attachment to cell membranes (2). A more rigorous experiment determined that apoJ also blocks complement C9 assembly on C5b-8 and C5b-9 (23).
A later study demonstrated that exogenous apoJ, at or below physiologic concentrations, provided no protective effect against serum-derived MAC activity in cell culture, and showed that cell-surface localized apoJ did not inhibit complement-mediated lysis (182).
In contrast to these variable results in vitro, studies ex vivo and in vivo suggest that apoJ suppresses immune-mediated disease. In the complement-dependent isolated perfused rat kidney injury model, serum depletion of apoJ augments immune glomerular injury (24). The apoJ- dependent injury is associated with elevated proteinuria and increased glomerular deposition of complement compared to controls. These findings were corroborated by a clinical study indicating that serum apoJ levels are inversely correlated with disease activity in SLE patients
(42).
Despite these numerous studies and observations, the physiologic function of apoJ in immune-mediated renal disease remains unclear. Besides complement regulation, it has been hypothesized that apoJ preserves tubular epithelial structure via the maintenance of cell-cell and/or cell-matrix interactions (9, 89). We propose that apoJ is a non-specific extracellular chaperone that mediates the clearance and disposal of immunogenic molecules generated at
108 tissue injury sites, thereby limiting inflammation. We also speculate that its chaperone function has evolved to regulate both humoral and cellular immune factors.
We recently showed that apoJ deficiency increases the severity of murine autoimmune myocarditis (43). At the disease onset, wild-type (WT) and apoJ-deficient (KO) mice present with equivalent responses, based on inflammatory markers and antibody production. Similar numbers of WT and KO animals develop pathologic lesions, but the extent of disease is significantly greater in the absence of apoJ, especially in female mice. This observation is associated with an increased incidence of cardiac failure in treated KO mice compared to WT mice. Interestingly, a robust humoral response against cardiac antigens in WT mice is associated with minimal disease, while a comparable response in KO mice is linked to severe disease. We postulated that apoJ suppresses autoimmune disease by regulating immune responses against self-antigens.
We have also observed that apoJ deficiency predisposes a progressive glomerulopathy in mice characterized by the hypocellular expansion of the mesangium with collapse of capillary lumens (Rosenberg et al., in preparation). In contrast to WT animals, KO animals accumulate IC deposits in the mesangium as early as 4 weeks of life. Immunohistochemistry demonstrates that the deposits contain IgG, IgM, IgA, and in some cases complement C3. The ICs are subsequently organized into dense mesangial deposits at 6 months of age, and by 21 months of age the ICs are arranged into tubulo-fibrillary structures. These observations imply that apoJ is an essential factor for the clearance of IC deposits.
Based on the differences between KO and WT animals in our myocarditis and glomerulopathy studies, I postulated that apoJ suppresses tissue damage in ICGN. I examined the role of apoJ in the horse spleen apoferritin (HAF) model of ICGN. The HAF model is a
109 well-characterized murine model of renal disease, and has been exploited to reveal numerous facets of ICGN. WT mice treated with HAF develop glomerular deposits that contain IgG, IgM,
IgA, and complement C3 (183). In contrast, complement C3- and complement C4-deficient animals generate deposits lacking IgG, but including elevated levels of IgM and IgA compared to WT animals. In addition, complement C4- deficient mice have greater deposition of complement C3 and subsequent increased glomerular hypercellularity compared to WT and complement C3-deficient animals. It has also been demonstrated that complement C5 is required for glomerular IC localization and the induction of progressive GN (184). Besides the complement pathway, a series of mouse experiments, examining interstrain variations in susceptibility to HAF-mediated disease, have implicated an essential role for the major histocompatibility complex in ICGN (185-187).
Using the HAF model, I found that apoJ-deficient animals consistently presented with more disease than WT animals in four independent trials. KO female mice were particularly sensitive to the kidney injury model, as characterized by hematuria and glomerular pathology.
IC composition and deposition were equivalent in both genotype backgrounds. However, the
Th2/Th1 immune response ratio was greater in KO mice than WT mice based on serum antibody profiles. Thus, apoJ-deficiency altered the normal immune response. These experimental results were very similar to those of our autoimmune myocarditis study and demonstrated that apoJ reduces inflammation and protects tissues against IC-mediated disease. The results also highlighted the role of gender bias in disease processes. In view of its diverse biological properties, I hypothesize that apoJ limits tissue injury by modulating the interactions between immune complexes and key effector molecules, such as complement proteins and Fc receptors.
110 MATERIALS AND METHODS
Generation of apoJ-deficient mice. The apoJ-deficient mice were generated by Guang
Zhu as described in McLaughin et al. (43). The mutant mice were first generated in the Swiss
Black outbred genetic background and subsequently backcrossed for seven generations into glomerulonephritis-susceptible strain Sv129 mice. Animals were housed in a pathogen free- barrier facility until the beginning of the experiment, at which time they were moved into a conventional non-pathogen free facility. ApoJ-deficient and wild-type 129 mice were obtained from the mating of heterozygous apoJ+/- animals, and from the mating of homozygous apoJ +/+ and -/- animals. Genotypic analysis of apoJ gene-targeted mice was carried out as indicated previously (43).
Apoferritin-mediated glomerulonephritis. (Trials 1-3): Six to 8 week old apoJ- deficient and wild-type strain Sv129 mice received intraperitoneal injections of 4 mg of horse spleen apoferritin (HAF) (Sigma Chemical Co., St. Louis, Missouri, USA, trials 1, 2; Calzyme
Laboratories, Inc., San Luis Obispo, California, USA, trial 3) in saline, 5 days a week, for 5 weeks (188). The mice were also injected with 50 µg of bacterial lipopolysaccharide from
Salmonella minnesota (Calbiochem-Novabiochem Intl., San Diego, California, USA) in saline, 3 days a week for 5 weeks. At times 0 and 5 weeks, animal mass was recorded, and an 12-hour urine specimen was collected in a metabolic cage. In the metabolic cage, mice had access to water, but not to food. At time 3 weeks, blood was obtained from the retroorbital venous plexus into heparinized capillary tubes, and the serum was isolated. At the conclusion of five weeks, the animals were sacrificed by carbon dioxide asphyxiation, and the blood was collected via
111 cardiac puncture. The kidneys were harvested, and their masses were determined. (Trial 4):
HAF (Sigma Chemical Co.) was emulsified in complete Freund’s adjuvant (CFA) (Gibco BRL,
Rockville, MD, USA), and 2 mg were injected subcutaneously. Ten days later, HAF was administered via intraperitoneal injections, at a dose of 4 mg, 5 days a week for 5 weeks
(females) or 7 weeks (males). Seven weeks after the primary immunization, the animals also began to receive 50 µg of LPS, 3 days a week, for the remainder of the experiment. Female and male mice were sacrificed eight and ten weeks following the primary immunization, respectively. Tissue and blood specimens were harvested as described above.
Blood urea nitrogen (BUN) detection. BUN levels were determined using the BUN acid and color reagents derived from the Sigma Chemical Co. Reactions containing 5 µl serum,
750 µl BUN acid reagent, and 500 µl BUN color reagent were incubated at 96°C for 10 minutes.
The samples were chilled in a cold water bath for 5 minutes, and then the colorimetric change in absorbance was measured at 525 nm. BUN concentrations were calculated using standard curve data.
Urinalysis. Total urinary protein was measured via the Bio-Rad protein assay (Bio-Rad
Laboratories Inc., Hercules, CA, USA) at 595 nm, based on the method of Bradford. Urinary creatinine levels were determined by the addition of picric acid under alkaline conditions to form
Janovski complexes. The complexes were detected at 500 nm using the Beckman creatinine analyzer 2 (Beckman Coulter, Inc., Fullerton, CA, USA). The extent of hematuria was determined using Multistix 10 SG reagent strips for urinalysis (Bayer Corporation, Elkart,
Indiana, USA).
112
Histology. Renal tissue was preserved in formalin, and 5-µm sections were cut and
stained with hematoxylin and eosin (H&E). Fifty randomly selected glomeruli per animal
section were scored on a scale of 0-4, with 0 = no glomerular lesion; 1 = segmental hypercellular
lesion; 2 = extensive hypercellularity throughout the glomerulus; 3 = a necrotizing lesion defined by the presence of karyorrhexis; 4 = crescentic nephritis.
Immunofluorescence studies. Renal tissue specimens for immunohistochemistry were initially preserved in isopentane (Sigma Chemical Co.) and stored at -70°C. At the time of analysis, 4-µm thick cryostat sections were cut, applied to slides, fixed in acetone, air dried, and washed in PBS. Fluorescein-5-isothiocyanate (FITC)-conjugated goat anti-mouse antibodies
(anti-IgG, IgM, IgA, complement C3, and complement C9; (ICN Pharmaceuticals, Inc., Costa
Mesa, CA, USA) were diluted 1:100 in PBS, and incubated with the tissue slides for 1 hour in a humidified chamber. The slides were washed in PBS, sealed with antibody mounting media and a coverslip, and stored at 4 °C prior to imaging.
Electron microscopy. Renal tissue specimens for electron microscopy were finely diced, fixed in osmium tetroxide for approximately 1 hour, and incubated in cacodylate buffer
prior to being embedded in KX112 resin, sectioned, and stained with 2% uranyl acetate with
ReynlB (lead citrate), or silver nitrate.
Western blot analysis of serum complement C3. Two hundred fifty nanograms of total
serum protein, diluted in SDS reducing buffer, were subjected to SDS-polyacrylamide gel
113 electrophoresis (PAGE) at 200 V for 1 hour on a 7.5% mini-gel. The proteins were transferred overnight at 4°C onto a polyvinylidene fluoride (PVDF) Immobilon-P membrane (Millipore Inc.,
Bedford, Massachusetts, USA). The membrane was treated with a 0.2% casein blocking solution
(Sigma Chemical Co.) solution, prepared in 0.1%Tween-20/Tris-buffered saline (TBS), for one
hour at room temperature with gentle rocking. The primary antibody, goat anti-mouse C3 IgG
(ICN Pharmaceuticals, Inc), was diluted 1:1600 in Tween20/TBS, and incubated with the
membrane at room temperature for 30 minutes, with gentle rocking. The membrane was
subsequently washed with Tween20/TBS, and incubated with a 1:1000 dilution of the secondary
antibody, biotinylated rabbit anti-goat IgG (Vector Laboratories Inc., Burlingame, CA, USA) for
30 minutes at room temperature with gentle rocking. Following an additional set of washes, the
membrane was developed with the ECL western blotting analysis system (Amersham Pharmacia
Biotech Inc., Piscataway, NJ, USA), and exposed to autoradiographic film. The film was
scanned and analyzed via ImageQuant software (Molecular Dynamics).
Anti-apoferritin, total IgG1, and total IgG2a ELISA. (Anti-apoferritin): Polystyrene
plates were treated with HAF (Sigma), diluted to a concentration of 10 µg/mL in binding buffer
(15 mM Na2CO3, 35 mM NaHCO3, 0.02% sodium azide, pH 9.5), and allowed to stand overnight at 4°C. The plates were washed 3 times with PBS-T (phosphate buffered saline with
0.05% Tween 20), and coated with blocking buffer (borate buffered saline, 0.05 % Tween 20, 1 mM EDTA, and 0.25% bovine serum albumin) for 1 hour at room temperature. The plates were washed 3 times with PBS-T, and then serum samples, diluted 1:50 and 1:250 in blocking buffer, were administered. Following an overnight incubation at 4°C, the plates were washed 3 times with PBS-T. Biotinylated secondary antibodies, anti-mouse IgG1 and IgG2a (Pharmingen), were
114 diluted in blocking buffer to a concentration of 2 µg/mL, and dispensed onto the plates for 1 hour at room temperature. The plates were washed six times with PBS-T, and treated with avidin peroxidase (Sigma Chemical Co.), diluted 1:400 in blocking buffer, for 30 minutes, at room temperature. After six additional washes with PBS-T, the plates were treated with the peroxidase substrate ATBS (2,2-azino-bis (3-ethylbenzthiazoneline-6suflfonic acid)) (Kirkegard Perry
Laboratories, Maryland, USA). Enzymatic development of the substrate was detected at 405 nm using a Lab Systems Multiscan EX machine. (Total IgG1 and total IgG2a): The assays were carried out as described above with the following changes: (a) the plates were initially coated with 2 µg/ml of anti-mouse IgG1 or anti-mouse IgG2a (Pharmingen) in PBS, and (b) the blocking solution was composed of 10% fetal bovine serum in PBS.
Statistical analyses. Chi Square analysis was utilized to evaluate the differences in the glomerular pathologic frequencies between animals groups per experimental trial. T-test analysis was used to ascertain the differences in mean serum antibody levels between animal groups per experimental trial.
115 RESULTS
Experimental model optimization. Though chronic HAF treatment is a well-
characterized IC-mediated disease model in rodents, previous investigators have described renal
disease variability based on strain genetic differences (185-187). I initially sought to optimize
the model by comparing the injury sensitivity of C57/B6, Sv129, and FVB/N wild-type mouse
strains. My preliminary studies demonstrated that FVB/N mice, which lack complement protein
C5, were resistant to HAF-mediated renal injury (data not shown). This observation correlated
with a prior study that stated the HAF model requires an intact complement system to induce
disease (184). In contrast, the Sv129 and C57/B6 mice developed moderate and severe
glomerular disease, respectively (data not shown).
I chose the Sv129 mouse strain for all further studies, as I felt the less extreme disease
progression would permit greater differentiation between a KO versus WT phenotype. I carried
out four independent trials to generate a statistically significant body of work. To stimulate
greater renal injury and overcome experimental inconsistency, I made changes in the treatment
regimen between trials. Trials 1 through 3 involved the chronic administration of HAF and lipopolysacchride (LPS) over a five-week period. In contrast, I initially immunized the animals in trial 4 with HAF in complete Freund’s adjuvant (CFA), and then after 10 days I began chronic administration of HAF for 5 to 7 weeks. I also included LPS treatment for the final 1 to 2 weeks of trial 4.
Histologic examination. Our primary assessment of renal disease in the treated animals involved the histologic analysis of glomerular pathology. The glomerular pathology was
116 quantified for each animal specimen using a 0 to 4 scoring system per glomerulus, in which 0 symbolized no pathology and 4 represented crescentic glomerulonephritis (Figure 1). WT and
KO kidney specimens from control animals were identical in appearance (data not shown). In response to IC-mediated injury, WT animals developed minimal renal disease, characterized by either no or segmental hypercellular glomerular lesions (Table 1). ApoJ-deficient mice presented with more progressive renal disease, symbolized by a higher occurrence of proliferative and crescentic glomeruli.
The extent of tissue injury varied between trials, as trial 1 animals developed significantly more renal pathology than animals in the other three trials, as symbolized by a higher percentage of crescentic glomeruli (Table 1). For trials 2-4, the majority of pathologic glomeruli were predominantly limited to grades 1 and 2. However, despite animal and trial variability, the KO and WT animal groups were statistically distinguishable in each of the four
trials based on histopathology. KO mice consistently had a greater percentage of pathologic
glomeruli than WT mice. There was also a gender bias observed in each experiment: KO and
WT female mice had higher pathologic scores than their respective male counterparts.
Physiologic analysis. To determine renal function and integrity, BUN levels were
measured, and urine specimens were analyzed for proteinuria and hematuria (Table 2). Untreated
KO and WT mice had equivalent BUN, urinary protein, and urinary blood quantities within
approximate normal limits (BUN ≈ 20 mg/dl, urinary protein/creatinine ratio ≈ 1, hematuria = 0)
(data not shown). Renal failure-related changes in BUN levels and proteinuria were only
detected in trial 1, and KO mice had dramatically higher concentrations of both BUN and urine
117 protein than WT mice. Neither BUN nor urinary protein amounts were significantly elevated in
the other three trials, which was consistent with the limited renal pathology.
Hence, most mice in my experiments were sub-clinical for renal failure, and more
similar to early human renal lesions. Hematuria proved to be the best renal injury indicator, as
urinary blood levels often directly correlated with glomerular pathology (data not shown). More than half of the treated KO female mice presented with hematuria. In contrast, animals from the three other groups infrequently developed hematuria except in association with severe disease.
In correlation with the lack of renal failure, the majority of the experimental animals survived the full treatment regimen for each trial. It should be noted though that none of the KO females survived the first experiment.
Immunofluorescence and electron microscopic analyses. ApoJ has been hypothesized to modulate IC formation and clearance, based on its known interactions with immunoglobulins
(119, 120). Though this hypothesis was not directly tested in my study, in trial 1 we qualitatively examined glomerular IC composition via immunofluorescence (IF) for IgG, IgM, IgA, complement C3, and complement C9. IgM deposition was strong but equivalent in WT and KO treated glomeruli (Figure 2). The IgM deposits were primarily localized to the mesangium, with some glomerular basement membrane (GBM) deposition. Moderate amounts of IgG were deposited in a similar manner as IgM in WT and KO treated mice, with possibly more GBM deposition. IgA glomeruli deposits were present in both genotypes, but to a much lesser extent.
KO and WT animals had abundant but comparable granular complement C3 deposits, predominantly localized along the GBM (Figure 3). The deposition of complement C9 was not
118 as strong as complement C3 in both animal groups. However, there appeared to be a limited
number of intense complement C9 focal deposits only in KO glomeruli.
Electron microscopic (EM) analyses also demonstrated similar glomerular IC deposits in
WT and KO treated animals derived from trials 1 and 2 (Figure 4). The deposits in the WT and
KO glomeruli appeared to be of the same size, and were principally localized to the mesangium, with occasional subepithelial and subendothelial deposition. Thus, the ICs in WT and KO animals were virtually indistinguishable by composition and glomerular location based on IF and
EM analyses. This observation suggests that apoJ restricts disease progression subsequent to IC formation and localization.
EM images also illustrated the abundance of infiltrating neutrophils in glomerular lesions, which correlates with an active inflammatory process (Figure 5). In addition, glomerular basement membrane ruptures were present in crescentic glomeruli. Basement membrane ruptures are pathognomonic for rapidly progressive GN (176). The degrees of neutrophil infiltration and basement membrane injury in WT and KO animals were not quantified.
Serum complement C3 levels. To determine if apoJ deficiency enhanced complement activity during glomerulonephritis, I measured serum complement C3 levels at the conclusions of trials 1 and 3 (Figure 6). The consumption of serum complement C3 is an indicator of severe disease for several human glomerulopathies, including SLE, post-streptococcal GN, and membranoproliferative GN (176). Serum complement C3 levels were comparable in WT and
KO control animals, and were relatively unchanged in response to HAF treatment. However, serum complement C3 concentrations in KO treated mice were slightly higher than in their WT
119 counterparts. In addition, serum complement C3 amounts in treated female mice were slightly
higher than in treated male mice. Neither difference was statistically significant.
Serum anti-HAF and total antibody levels. I measured both anti-HAF and total serum
antibody levels at the conclusions of trials 2-4 to establish if apoJ deficiency altered the immune
response (Figures 7 and 8). Special consideration was given to Th1-specific IgG2a and Th2-
specific IgG1 antibody isotypes in order to differentiate between Th1 versus Th2 immune
responses. Prior to treatment, anti-HAF serum antibodies were absent in representative KO and
WT animals (relative antibody levels for WT control normalized to 1). Total serum antibody
levels were equally low in both animal groups (relative antibody levels for WT control normalized to 1).
In response to HAF treatment, KO and WT animals had increased, but equivalent amounts of anti-HAF IgG1 and total IgG1 serum antibodies in trials 2, 3, and 4, and anti-HAF
IgG2a serum antibodies in trials 2 and 3. In contrast, total IgG2a antibody amounts were statistically higher in WT mice than in KO mice in trials 2 and 3. Trial 4 antibody profiles differed from the other trials, as there was no HAF-dependent elevation in anti-IgG2a serum antibody levels, nor a genotype-dependent difference in total IgG2a serum antibodies.
Based on the ratio of total IgG1 to total IgG2a total serum antibody levels, HAF administration induced a polarized Th2 immune response in all experimental animals (Figure 9).
In addition, the generalized Th2 immune imbalance was statistically greater in KO animals versus WT animals in trials 2 and 3. However, the polarized Th2 response was not observed at the level of anti-HAF antibodies in either WT or KO animals.
120 Experimental variability. Experimental inconsistency is a major challenge in all animal
models of disease, and is predisposed by a variety of confounding variables including animal
housing, strain susceptibility, and reagent preparation. Litter to litter variability regarding
glomerular pathology was common in each trial of my study (data not shown), and has been
observed in the HAF model by other investigators (189). Trial to trial variability concerning the
extent of disease was also significant. For example, trial 1 animals presented with severe disease
that was unmatched by animals in the other three trials. However, in spite of experimental
inconsistencies regarding treatment regimen and disease progression, KO mice were always
more sensitive than WT mice to HAF-induced renal injury. I believe this observation, in the
light of the various confounding variables, strengthens the arguments for the pathophysiologic
importance of apoJ.
I suspect that the use of different stock solutions of HAF and LPS was the primary cause
for the trial inconsistency. The reagent changes may have been particularly important for HAF,
as I treated the animals in trials 1, 2, and 4 with an 85% pure HAF solution prepared by Sigma
Chemical Corporation (St. Louis, Missouri, USA). The 15% of impurities in the Sigma HAF
preparation confers bioactive differences between lots, and thus may have dramatically altered disease progression. For trial 3, I introduced a potentially important confounding variable in the study by changing the source of HAF from Sigma to Calzyme Laboratories, Inc. (San Luis
Obispo, California, USA). In addition, the purity of the HAF from Calzyme was approximately
95%. The treatment regimen of trial 4 was completely different than the other 3 trials based on
the inclusion of CFA and the extended duration of the trial. Therefore, trial 4 data should not be compared to data from the other three trials.
121 DISCUSSION
ApoJ has long been associated with tissue injury and disease pathogenesis, including
many renal disorders. Though speculated to be a protective molecule, its exact function (or
functions) remains unknown. My work demonstrated, for the first time in vivo, that apoJ limits
ICGN. In combination with our earlier findings relating to autoimmune myocarditis, my
observations suggest that apoJ is a general anti-inflammatory factor that suppresses multiple
types of immune-dependent tissue injury.
The results from four independent trials demonstrated that apoJ deficiency augments IC-
mediated kidney disease. Following the chronic administration of HAF, KO animals
consistently presented with more glomerular pathology than their WT counterparts in spite of
experimental variability. The enhanced renal disease in the KO genetic background was often
accompanied by hematuria, and in some cases elevated BUN and proteinuria. Besides apoJ
genotype, the role of gender bias was also apparent in my model system, as WT and KO female mice developed more severe disease than genotype equivalent male mice. Consequently, KO female animals were predisposed to glomerular injury.
Postulated mechanisms of apoJ-mediated suppression of IC-mediated disease. The
pathogenesis of ICGN is a complex sequential process that involves numerous soluble and
cellular mediators (Figure 10) (176, 177). ICGN is promoted by the Type III hypersensitivity
reaction, classically characterized by serum sickness. Antibodies against an antigen, self or foreign, are generated, and subsequently form ICs. The ICs are deposited in renal glomeruli, where they then activate complement proteins and attract multiple inflammatory cell types,
122 including neutrophils, platelets, and macrophages. Complement proteins recruit additional
inflammatory cells, and mediate tissue injury via MAC-dependent cell lysis. Inflammatory cells,
as well as endogenous renal cells injured by the MAC, release toxic agents such as oxidants and
proteases that promote further tissue damage. The combined effects of these various factors
predispose the kidney to proteinuria, cellular proliferation, and renal failure.
Considering the generality of apoJ interactions, it is probable that apoJ has multiple functions in response to IC-mediated disease. For example, via its chaperone activity, apoJ may
bind ICs, change their three-dimensional structure, and subsequently reduce IC antigenicity.
ApoJ may sequester and inhibit pro-inflammatory factors such as chemokines and cytokines.
Additionally, apoJ may clear and dispose of toxic inflammatory and apoptotic byproducts
including oxidized lipids and proteases. Based on its ability to facilitate cell aggregation, apoJ
may maintain cell-cell and cell-matrix interactions during tissue destruction. Conversely, apoJ
may transport inflammatory cells and/or dead cells away from a pathologic lesion. Thus, apoJ
has the biological potential to affect both the initiation and resolution of tissue injury in
numerous ways.
However, based on my data and those of others, I propose that apoJ suppresses ICGN by
1) limiting the binding of ICs to complement proteins and cellular receptors, and 2) suppressing
the MAC. By binding immunoglobulins in vivo, apoJ may block interactions between ICs and
the first component of the complement cascade. ICs activate the classical complement pathway
by binding complement C1 (176, 190). C1 in turn activates complement C2 and complement
C4. As with the other complement factors, the non-specific activation of C1 is limited by a
regulatory protein, namely C1-inhibitor (C1-In) (190). C1-In, a serum factor, is essential for
controlling the spontaneous activation of C1, and its deficiency predisposes the human disease
123 hereditary angioedema. In contrast, C1-In has little effect on IC-induced activation of C1, and
accordingly, is only rarely associated with IC-mediated disease (190-194). Therefore, apoJ may
be the functional equivalent of C1-In in response to IC-C1 interactions. This idea is especially
attractive considering C1q, a C1 subunit, augments apoJ-IC binding (195).
ApoJ may also block the interactions between ICs and cellular receptors for the Fc
domain of IgG antibodies (FcRs). FcRs provide an essential link between humoral and cellular
immune responses, and are required for type II and type III hypersensitivity (196, 197). FcRs are
found on the surface of multiple cell types including leukocytes and mesangial cells, and their
stimulation induces antibody dependent cytotoxicity (ADCC), phagocytosis, and the secretion of
pro-inflammatory mediators (196, 198, 199). FcRs also modulate antibody production and IC clearance (196). Thus, FcRs both facilitate and inhibit immune effector pathways.
Recent studies have demonstrated the biological significance of FcRs in IC-mediated
renal disease. IC treatment of cultured mesangial cells activates FcRs and promotes cellular
proliferation, matrix synthesis, and chemokine/cytokine release (198). Mice deficient in the FcR
gamma (γ) subunit are resistant to nephrotoxic GN and various lupus models (200, 201). In
addition, peptides that block IC-FcR interactions, such as the Fc fragment itself, inhibit IC-
dependent renal injury (198, 202). These results indicate that FcRs are a viable target for the
treatment of human autoimmune diseases (196).
ICs, via their interactions with FcRs, affect the balance between Th1 and Th2 immune
responses (203-206) (Figure 11). Th1 and Th2 immune responses are directed by different
subsets of helper (CD4) T cells, and are characterized by distinct antibody and cytokine profiles
(207). Th1 cells promote cellular immunity, exemplified by type IV hypersensitivity (DTH), the
destruction of intracellular pathogens, and organ-specific autoimmune diseases. Th1 immune
124 responses are characterized by the synthesis of IgG1 and IgG3 antibodies in humans, and IgG2a
and IgG3 antibodies in rodents. Th1 cells secrete the cytokine interleukin-2 (Il-2), which induces
inflammation by activating all lymphocytes (208). Th1 cells also stimulate macrophage killing
and phagocytosis via the secretion of interferon gamma (IFNγ). In turn, macrophages synthesize
interleukin-12 (Il-12) that stimulates Th1 cell growth.
Th2 cells promote humoral immune responses, such as allergic reactions or type I
hypersensitivity, and systemic autoimmune diseases (207). IgG4 antibodies in humans and IgG1 antibodies in rodents characterize Th2 immune responses. The primary Th2 specific cytokine is interleukin-4 (Il-4). Il-4 stimulates B cell growth, isotype switching, and IgE antibody synthesis
(208). Additionally, Il-4, along with another Th2 cytokine, interleukin-10 (Il-10), has the capability of converting a Th1 immune response into a Th2 response by increasing the growth of
Th2 cells and inhibiting the growth of Th1 cells, respectively (207, 208). Conversely, IFNγ prevents Th2 cell development, and inhibits Il-4 dependent induction of B cell proliferation and differentiation. As a result, Th1 and Th2 immune responses are tightly regulated via interpathway feedback loops. In the absence of proper regulation, the shift toward either immune response predisposes disease.
An immune imbalance in favor of a Th2 response is associated with various systemic autoimmune disorders, including SLE. SLE patients have increased gene expression and serum levels of multiple Th2 cytokines, including Il-4 and Il-10, compared to control patients (209-
211). In addition, Th2 cytokine levels directly correlate with disease activity, especially at the onset of symptoms (209, 212-214). In contrast, production of Th1 cytokines, such as IFNγ and
Il-12, is either absent or repressed in many SLE cases (209-211, 215). These findings demonstrate a profound imbalance or skewing towards a polar Th2 immune response in SLE
125 pathogenesis. Consequently, pharmacologic suppression of SLE and immune complex
glomerulonephritis is associated with a Th2 to a Th1 immune response shift, symbolized by the
induction of Il-12 and IFNγ, and the repression of Il-4 and Il-10 (216-218).
Th2 immune responses have also been implicated in specific renal injuries. Membranous
GN involves the deposition of IgG4, a Th2 specific antibody (207). Minimal change GN may be
caused by systemic immune stimulation, and is associated with enhanced synthesis of Il-4 (219).
In experimental models, Th2-prone mice are more sensitive than Th1-prone mice to anti-
glomerular basement membrane antibodies (220). Upon administration of such antibodies, Th2-
prone mice develop acute glomerulonephritis characterized by severe proteinuria and glomerular
deposition of immunoglobulin and complement. On the contrary, Th1 mice present with mild
proteinuria and limited glomerular immune complex deposition.
In my study, total Th2 serum antibody levels were greater than total Th1 serum antibody levels in all treated animals, suggesting that the HAF-induced GN is mediated by a Th2- dependent mechanism. The generalized Th2 immune response in my model system is probably enhanced by the inclusion of LPS, a well-characterized polyclonal B cell activator. Despite the
Th2 predominance, WT mice developed a moderate Th1 immune response in trials 2 and 3. In contrast, KO animals developed a minimal Th1 immune response, and hence presented with an exceptionally polarized Th2 phenotype. This immune imbalance could be either a cause or consequence of the increased inflammatory reaction in the KO animals.
I believe the skewed Th2 serum antibody profile in KO mice represents the biological origin of their increased sensitivity to HAF-induced renal injury. The elevated Th1 immune response in WT animals may protect against the Th2-dependent ICGN. Hence, the lack of a protective Th1 immune response in KO animals may predispose progressive disease. I
126 hypothesize that apoJ deficiency augments a Th1 to Th2 immune response shift that enhances
glomerular pathology in the HAF model system (Figure 11). This theory is attractive
considering the inverse relationship in humans between serum levels of apoJ and Th2-dependent
SLE disease activity (42).
ApoJ deficiency may increase Th2 polarization and humoral-dependent disease by
enhancing IC-FcR interactions. ICs induce a Th1 to Th2 shift in vitro by stimulating the
secretion of Il-10 and inhibiting the secretion of Il-12 (203-206). By binding immunoglobulins,
apoJ may limit IC-FcR interactions in vivo, thus directly repressing FcR activation. This action
would indirectly block Th2 cell induction, and in turn may promote Th1 cell induction. On the
other hand, apoJ may facilitate protein-protein and/or cell-protein interactions that directly
stimulate the Th1 pathway, which would counteract Th2-dependent activities.
It should be emphasized that the Th2 immune polarization did not involve the initial
reaction against HAF, as Th1 and Th2 anti-HAF serum antibody levels were elevated, but
equivalent in most treated animals. In addition, serum concentrations of the anti-HAF antibodies
were identical in the WT and KO treated animals. Therefore apoJ is not required to mount a primary humoral response against HAF. This evidence is consistent with the IF and EM data, and suggests that the apoJ-dependent renal injury is subsequent to immune complex deposition.
The significance of the Th2 immune response in my study is further strengthened by the observation of female sensitivity. Both WT and KO female mice presented with more severe disease than their male counterparts. This finding is consistent with many other autoimmune diseases and experimental models. SLE, scleroderma, and Sjogren’s syndrome are particularly prevalent in females (176). We also observed a female bias in our myocarditis study (43).
127 Accordingly, female mice proved to be very powerful in my experiments for highlighting the
apoJ-deficient phenotype.
Hormonal differences are a common theory for the gender bias in autoimmune disorders.
For example, SLE disease activity increases in response to pregnancy-induced changes in estrogen levels (221, 222). The elevated inflammation directly correlates with a Th1 to a Th2 immune response shift based on cytokine expression profiles. Under similar conditions, Th1 dependent diseases, like rheumatoid arthritis, are suppressed. Estrogen induces polyclonal B-cell activation, and consequently augments lupus nephritis (223). Conversely, inhibition of estrogen activity via tamoxifen treatment suppresses ICGN (224). It has also been observed that T cells derived from females innately synthesize lower levels of Il-2 and IFNγ and higher levels of Il-4 and Il-10 than male T cells, thus predisposing Th2 polarization and autoimmune disease (225).
ApoJ-dependent suppression of ICGN also likely involves complement regulation since apoJ is an invariable component of ICs containing SC5b-9, and is a potent inhibitor of complement-dependent cytotoxicity in vitro (2, 181). Repression of complement activity is a significant therapeutic avenue for numerous types of immune injury in vivo. As previously mentioned, complement C5 deficiency ameliorates HAF-induced GN, and administration of a blocking anti-complement C5 antibody attenuates disease in a murine lupus model (184, 226).
Additionally, rats deficient in complement C6 do not form the MAC, and are subsequently protected from the anti-thymocyte serum (ATS) model of mesangioproliferative GN (227).
To limit non-specific or ectopic complement activity, mammals have evolved an extensive repertoire of membrane-bound complement regulatory proteins. The regulatory factors include membrane cofactor protein (MCP), complement receptor 1 (CR1), and delay accelerating factor (DAF), all of which are co-factors involved in the decay of the C4b and/or C3b subunits of
128 the C3/C5 convertases (228). An additional factor, CD59, blocks the synthesis of the MAC.
Changes in the expression of these various molecules have been observed in immune-mediated
disease. For example in the normal kidney, DAF synthesis is restricted to the juxtamedullary
apparatus and a few glomerular cells (229). However, in cases of IgA nephropathy (IgAN),
membranous nephropathy, and lupus nephritis (LN), DAF is broadly expressed by tubular and
mesangial cells. DAF expression directly correlates with disease activity in IgAN, and in some
cases of LN (230). The glomerular synthesis of CD59 is also included in LN. In contrast, CR1
glomeruli expression is repressed in LN.
Experimental models have demonstrated that the alteration of complement regulation can
modulate IC-mediated tissue injury. Anti-CD59 blocking antibodies increase the glomerular deposition of the MAC, and predispose severe endothelial cell injury and renal failure in a rat model of immune thrombotic microangiopathy (TMA) (231). In addition, a monoclonal blocking antibody against CD59 enhances glomerulonephritis in an alternate rat model (232).
Conversely, the chronic administration or transgenic over-expression of complement receptor 1– related gene/protein y (Crry), a rodent homologue of CR1, suppresses nephrotoxic serum (NTS) induced nephritis in mice (236.)
Complement activity was not quantitatively assessed in my study. I did however measure serum complement C3 levels. Prior to the experiment, I speculated that serum C3 levels would be lower in KO animals than WT animals due to hypothetically enhanced complement deposition in the severely injured kidneys. In contrast, I observed that serum levels of complement C3 were essentially unaffected by the chronic administration of HAF compared to control samples. Interestingly, it appeared that serum C3 levels were directly correlated with disease activity, as KO and female mice had greater levels of serum C3 than WT and male mice,
129 respectively. Considering the use of LPS in my study, and the LPS-dependent induction of C3 in the acute phase response, elevated serum C3 levels in the more sensitive animals may be associated with an increased systemic inflammatory reaction (233-235).
Complement C3 and complement C9 glomerular deposition were also assessed, and the
IF results suggested that C3 was activated and deposited in a similar fashion in the WT and KO genetic background. The same could be said for C9, except for an occasional focal deposit of C9 in KO but not WT glomeruli. This observation is intriguing, based on the ability of apoJ to regulate MAC function in vitro, and may suggest that apoJ deficiency increases MAC deposition
in the HAF model system.
ApoJ may be a supplementary complement regulatory protein considering its in vitro and
in vivo complement related properties. The expression profile of apoJ is comparable to the other complement regulators in many disease states, implying a common protective function. For example apoJ and CD59 levels are minimal in control breast tissue, but both are significantly induced in breast carcinomas, possibly conferring tumor cell survival against immune attack.
Like DAF and CD59, apoJ is also upregulated in mesangial cells in response to some models of
ICGN (178). However, preliminary microarray analysis of WT HAF-treated kidneys indicated that, while apoJ expression was induced, DAF synthesis was actually repressed compared to WT control samples (personal communication, Bruce Aronow). Conversely, DAF synthesis was increased in KO treated animals compared to KO control animals (personal communication,
Bruce Aronow). These results suggest that a compensating balance of the regulatory factors is required to ensure a proper protective response during immune-mediated injury.
In contrast to the other complement regulatory factors, apoJ is considered to be primarily a secreted protein rather than a membrane bound factor. However, secreted apoJ has the
130 potential to bind a variety of lipid-based molecules including cellular membranes via its multiple
amphipathic helices. The association of apoJ with HDL particles may also facilitate complement
regulation at membrane surfaces, as HDL has an intrinsic complement inhibitory activity (34,
262). ApoJ is also structurally divergent from the other complement regulators. Nevertheless,
apoJ specifically blocks the formation and cytolytic activity of the MAC, much like CD59.
These similarities and dissimilarities between apoJ and the complement regulatory proteins
suggest that apoJ may represent a novel class of secreted complement regulatory factors.
Future work. My study is the first demonstration of a highly sought after role for apoJ
in ICGN in vivo, and thus represents an exciting step forward. However, the present study has
only highlighted the significance of apoJ, and now further experimentation is required to fully dissect the mechanism(s) of apoJ-mediated suppression of IC-dependent injury. Considering the extended experimental time, high costs, and disease variability related with HAF model system, I
would first strongly recommend switching to an alternative nephritis model, such as the
nephrotoxic serum (NTS) model system (236). In addition, it would be very interesting to
examine the effects of apoJ deficiency in a murine lupus model such as in the MRL-lpr mouse
strain (237). Based on my results utilizing the HAF model, I would expect apoJ deficiency to augment IC-induced pathology in any model.
Independent of the specific model system, numerous aspects of IC-induced disease need to be assessed in the absence and presence of apoJ in vivo. I would propose to carry out a rigorous time course study for each analysis to determine the exact timing of apoJ action during pathogenesis. To determine if apoJ facilitates IC disposal, IC complex clearance could be measured via the use of radioactive or fluorescent substrates. The three dimensional properties
131 of IC deposits could be assessed utilizing confocal microscopy to establish if apoJ modulates IC formation and structure, as well as their relationship with other inflammatory factors such as complement. In order to characterize the apoJ-dependent immune response, one could carry out a broad gene expression analysis by using microarray technology.
The tissue origin of apoJ in GN is a controversial topic, as conflicting studies have implicated both glomerular apoJ expression and systemic sources such as the serum and platelets
(116, 178). The relationship between platelets and GN is interesting considering the abundance of apoJ in platelet α-granules (60). Platelets enhance glomerular injury and inflammation by the discharge of histamine, serotonin, and eicosanoids (176). In addition, platelets release growth factors, such as transforming growth factor-β (TGF-β) and platelet derived growth factor
(PDGF), which accelerate glomerular fibrosis. It has also been demonstrated that platelet- inhibitory drugs minimize renal injury (238, 239).
Platelet-derived apoJ may counteract the various pro-inflammatory factors, and accordingly its deficiency would confer a disequilibrium favoring marked tissue injury.
Consistent with this theory, deficiency of platelet-derived apoJ may augment disease activity in
SLE. Depressed apoJ serum levels in active SLE patients are thought not to be caused by increased complement consumption or steroid use, but rather autoimmune–mediated platelet destruction (42). Consequently, a vicious cycle is generated by which additional serum levels of apoJ are depleted, thus promoting further tissue inflammation.
WT tissue-specific replacement therapy in the KO animal should be pursued to determine the source of apoJ in GN, as well as to verify the protective nature of apoJ. For example, WT serum-derived apoJ could be tested for its ability to confer GN resistance following injection into
KO animals. Via bone marrow replacement, the function of WT platelet-derived apoJ could be
132 dissected in KO mice challenged with IC-mediated disease. In addition, this experiment could
be carried out in reverse to investigate the role of non-platelet-derived apoJ in the absence of WT
platelets. In the same fashion, renal-derived apoJ could be studied by means of WT kidney
transplantation into a KO animal, and vice versa.
More complicated transgenic experiments could be carried out to further engage these
scientific questions. The use of a tissue-specific or conditional gene targeting strategy could be
powerful in determining the source and timing of apoJ activity. Conversely, a series of
constitutive tissue-specific apoJ transgenes could increase resistance against IC induced tissue injury in both KO animals and WT animals, while confirming the origin of the protective apoJ fraction. I also believe that dual gene targeting strategies including the apoJ gene locus could be utilized to either amplify or rescue the apoJ-deficient phenotype, thereby establishing the disease stage of apoJ action and any apoJ-interacting factors. Along the same lines, administration of blocking antibodies against various GN related factors could be examined in KO mice.
Clinical implications. The ability of apoJ to suppress ICGN in vivo has significant
implications for future patient care. As demonstrated in a SLE clinical trial, serum and/or tissue
levels of apoJ may represent a novel diagnostic and prognostic indicator of autoimmune disease
and systemic inflammation (42). Consequently, measuring the concentration of serum apoJ may
be useful in monitoring disease activity and therapeutic efficacy. Changes in serum apoJ may be
particularly important for detecting disease onset in newly presenting patients, or those patients
whose symptoms wax and wane. In addition, primary apoJ screenings may be useful in
predicting autoimmune disease potential in predisposed patient populations.
133 ApoJ replacement may become a therapeutic avenue for the treatment of autoimmune
disease and inflammation in general. The proposed animal experiments using WT fluids and tissues in challenged animals will be essential in determining the viability of apoJ therapy.
Since apoJ is abundant in most body fluids and has significant hydrophobic binding properties, it is probable that a pharmacologic preparation of apoJ would be stable upon administration, and would be delivered to most tissues and organ compartments. And in contrast to current autoimmune therapies, such as the steroidal and non-steroidal anti-inflammatory agents, systemic apoJ may be tolerated in large concentrations, over an extended period of time, with few adverse effects. Regardless of its future medicinal role, the study of apoJ function in human and experimental disease will further our understanding of all facets of pathogenesis, from the initial insult, to the final step in healing.
134 FIGURES AND TABLES
Figure 1. Scoring nomenclature for glomerular histopathology. High power magnification images of H&E stained glomeruli representing the 5 histopathologic scoring levels: 0 = no glomerular lesion; 1 = segmental hypercellular lesion; 2 = global hypercellular lesion; 3 = a necrotizing lesion defined by the presence of karyorrhexis; and 4 = crescentic nephritis. For reference, an asterisk (*) marks the center of each glomerulus.
135 0
*
* *
1 2
* *
3 4
Figure 1. Scoring nomenclature for glomerular histopathology.
136 Table 1. Histopathological assessment of renal sections. Fifty randomly selected glomeruli
per animal section from each experimental trial were scored based on a scale of 0-4 as described above. The experimental animals were divided into four groups based on genotype and gender, and the individual animal data was subsequently pooled per group per trial. The total number of pooled glomeruli under analysis per group per trial is cited in the third column. Values in columns 4 through 8 represent the percentages of glomeruli per pathologic score per group per trial. Chi square analysis indicated that the percentage of pathologic glomeruli was statistically greater in KO mice than WT mice for each trial (trial 1, males only, p < 0.0001; trials 2-4, males
and females combined, p < 0.0005). Chi square analysis also determined that the percentage of
pathologic glomeruli was statistically greater in female mice than genotype matched male mice
for trials 2-4 (trial 2, WT and KO mice, p < 0.025; trial 3, WT mice, p < 0.0005, KO mice, p <
0.005; trial 4, WT and KO mice, p < 0.0005).
137
Table 1. Histopathological assessment of renal sections.
% glomeruli per histopathologic score
Animal Group Trial total # 0 1 2 3 4 glomeruli
WT male 1 100 70 0 20 5 5 2 300 98 1 1 0 0 3 200 90 9 2 0 0 4 350 63 36 1 0 0
WT female 1 100 55 2 37 5 5 2 300 91 5 1 2 1 3 300 64 28 3 2 2 4 150 39 58 3 0 0
KO male 1 200 23 1 33 11 34 2 350 81 6 9 1 2 3 300 63 26 7 2 2 4 450 52 47 1 0 0
KO female 1 N/A 2 350 75 4 13 3 5 3 500 51 37 9 1 2 4 700 18 70 11 1 0
138
Table 2. Physiologic assessment of apoferrtin-treated mice. Multiple renal function parameters were measured at the experimental conclusion of each trial including serum blood urea nitrogen (BUN), urinary protein/urinary creatinine ratio (Pr/Cr), and presence of hematuria.
For BUN and Pr/Cr, each value symbolizes the mean per animal group per trial, +/- the standard mean error. Hematuria is reported as the number of positive urine samples per animal group per trial. The survival rate for each experimental group is indicated as the percentage of animals that survived the complete course of treatment. The “n” for each value is cited in parentheses.
139
Table 2. Physiologic assessment of apoferrtin-treated mice.
Animal Trial BUN Pr/Cr Hematuria % Survival Group (mg/dl)
WT male 1 28.5 ± 2.5 (2) 2.1 ± 0.8 (2) 1 (2) 67 (3) 2 17.0 ± 0.6 (5) 2.2 ± 0.3 (5) 2 (5) 86 (7) 3 22.2 ± 0.7 (4) 2.6 ± 0.6 (3) 0 (3) 80 (5) 4 21.1 ± 0.8 (6) 1.7 ± 0.2 (4) 1 (4) 75 (8)
WT female 1 29.0 ± 2.7 (3) 1 ± 0.5 (3) 0 (3) 100 (3) 2 22.9 ± 2.6 (5) 1.6 ± 0.4 (6) 2 (6) 100 (6) 3 28.9 ± 2.9 (6) 0.5 ± 0.2 (6) 2 (5) 100 (6) 4 31.0 ±0.4 (3) 1.0 ± 0.1 (2) 1 (4) 50 (8)
KO male 1 89.0 ± 30 (3) 4.4 ± 1.3 (4) 3 (4) 100 (4) 2 18.2 ± 1.5 (5) 2.9 ± 0.3 (7) 1 (7) 100 (7) 3 24.0 ± 3.2 (7) 1.7 ± 0.3 (7) 1 (7) 100 (7) 4 20.8 ± 1.1 (9) 1.6 ± 0.3 (7) 0 (7) 100 (9)
KO female 1 - - - 0 (3) 2 20.4 ± 2.0 (6) 1.7 ± 0.1 (7) 2 (7) 100 (7) 3 25.4 ± 1.8 (9) 1.4 ± 0.2 (9) 3 (9) 91 (11) 4 24.2 ± 1.7 (14) 1.5 ± 0.1 (3) 8 (9) 82 (17)
140 Figure 2. Immunohistochemistry for IgG, IgM, and IgA in representative glomeruli. Renal sections from representative trial 1wild-type (A, C, E) or apoJ-deficient (B, D, F) treated mice were examined for IgG (A, B), IgM (C, D), and IgA (E, F) glomerular deposition via immunohistochemistry.
141
A B
C D
E F
Figure 2. Immunohistochemistry for IgG, IgM, and IgA in representative glomeruli.
142 Figure 3. Immunohistochemistry for complement proteins C3 and C9 in representative glomeruli. Renal sections from representative trial 1 wild-type (A, C) or apoJ- deficient (B, D) treated mice were examined for C3 (A, B), and C9 (C, D) glomerular deposition via immunohistochemistry.
143
A B
C D
Figure 3. Immunohistochemistry for complement proteins C3 and C9 in representative glomeruli.
144 Figure 4. Electron microscopic analysis of immune complexes in representative glomeruli. Representative glomerular sections from wild-type (A, C) and apoJ-deficient (B, D)
treated animals were analyzed by electron microscopy. Immune complexes were detected
primarily in the mesangium (A, B), with some deposition in subepithelial and subendothelial
locations (C, D). For reference, immune complexes are noted with arrows. Images A and B
were captured at low magnification; images C and D were captured at a higher magnification.
All images were derived from trials 1 and 2 specimens.
145
A B
C D
Figure 4. Electron microscopic analysis of immune complexes in representative glomeruli.
146 Figure 5. Characteristic electron microscopic glomerular images. Glomerular image (A)
represents the abundance of vascular and urinary space in a control animal specimen. In
contrast, glomerular image (B) depicts characteristic hypercellularity in a treated animal specimen. Many of the cells are infiltrating neutrophils, marked by asterisks (*). Image (C) is derived from a crescentic glomerulus, and demonstrates a basement membrane rupture (flanked by the arrows). All images were derived from trials 1 and 2 apoJ-deficient specimens, and were captured at the same magnification.
147 A B
B C * *
*
Figure 5. Characteristic electron microscopic glomerular images.
148 Figure 6. Serum complement C3 levels in wild-type and apoJ-deficient treated mice in trials 1 and 3. Serum complement C3 levels were measured at the conclusions of trials 1 and 3 via a Western blot assay. The data points represent the mean fold change of serum C3 levels in treated animals versus pooled control samples per experimental group for both trials (WT♂ N=
6, WT♀ N = 8, KO♂ N= 9, KO♀ N = 8). The error bars symbolize the standard errors of the means.
149 2
1 treated:control Serum C3 levels
0 WT ♂ WT ♀ KO ♂ KO ♀
Figure 6. Serum complement C3 levels in wild-type and apoJ-deficient
treated mice in trials 1 and 3.
150 Figure 7. Anti-apoferritin IgG1 and total IgG1 serum antibody levels. Anti-apoferritin
(HAF) IgG1 (left panels) and total IgG1 (right panels) serum antibody levels were measured in
control and treated animal samples from trials 2, 3, and 4 using an ELISA protocol. Relative antibody levels were derived from colorimetric readings at 405 nm, and were normalized against the representative WT control sample per assay (anti-HAF or total) per trial. Average antibody levels are reported for the treated groups, along with the standard errors of the means. N = 1 for control groups (female), N = 9 for treated groups (5 females, 4 males).
151 Anti-HAF IgG1 Total IgG1 9 8 Trial 2 B 7 6 5 4 3 2 1 0 9 8 Trial 3 7 6 5 4 3 2
relative antibody levels relative antibody levels 1 0 9 8 Trial 4 WT control 7 KO control 6 WT treated 5 4 KO treated 3 2 1 0
Figure 7. Anti-apoferritin IgG1 and total IgG1 serum antibody levels.
152 Figure 8. Anti-apoferritin IgG2a and total IgG2a serum antibody levels. Anti-apoferritin
(HAF) IgG2a (left panels) and total IgG2a (right panels) serum antibody levels were measured in
control and treated animal samples from trials 2, 3, and 4 using an ELISA protocol. Relative antibody levels were derived from colorimetric readings at 405 nm, and were normalized against the representative WT control sample per assay (anti-HAF or total) per trial. Average antibody levels are reported for the treated groups, along with the standard errors of the means. N = 1 for control groups (female), N = 9 for treated groups (5 females, 4 males). T-test analysis indicated that the total IgG2a mean values were statistically higher in WT mice than in KO mice for trials
2 and 3 (trial 2, p < 0.02; trials 3, p = 0.05).
153 Anti-HAF IgG2a Total IgG2a 9 8 Trial 2 7 6 5 Series1 4 3 pp < < 0.02 0.02 2 1 0 9 8 Trial 3 7 6 5 Series1 p = 0.05 4 3 2
relative antibody levels relative antibody levels 1 0 9 8 Trial 4 WT control 7 KO control 6 WT treated 5 4 KO treated 3 2 1 0
Figure 8. Anti-apoferritin IgG2a and total IgG2a serum antibody levels.
154 Figure 9. Ratios of Th2 versus Th1 serum antibodies. The ratios of anti-apoferritin (HAF)
Th2 versus Th1 serum antibodies (top panel), and total Th2 versus Th1 serum antibodies (bottom panel) in treated animals were determined by dividing the respective ELISA values derived from identical serum specimens. The data points represent the average antibody ratios per animal group (WT vs. KO) per trial. The error bars symbolize the standard errors of the means. N = 9
(5 females, 4 males) per animal group per trial. T-test analysis indicated that the total Th2 versus
Th1 mean ratios were statistically higher in KO mice than in WT mice for trials 2 and 3 (trial 2, p = 0.004; trial 3, p = 0.003).
155 16 anti-HAF 14 WT treated KO treated 12
10
8
6
4
2
0 16 p = 0.003 Total 14
12
Th2:Th1 antibody level ratios Th2:Th1 antibody level 10
8
6 p = 0.004 4
2
0 Trial 2 Trial 3 Trial 4
Figure 9. Ratios of Th2 versus Th1 serum antibodies.
156 Figure 10. Postulated mechanisms of apoJ-mediated suppression of immune complex-
mediated glomerulonephritis. This diagram depicts the pathogenesis of immune complex-
mediated glomerulonephritis. Renal injury is promoted by non-inflammatory factors such as the
complement cascade, and a variety of inflammatory cell mediators including neutrophils,
platelets, macrophages, and mesangial cells. ApoJ suppresses glomerulonephritis, possibly by limiting the interactions between immune complexes and complement proteins, and immune complexes and the cellular receptors for the immunoglobulin IgG Fc domain (FcRs).
Additionally, apoJ may inhibit complement-mediated cell lysis. This drawing is copied and
adapted from the work of Couser, W.G. (176, 177).
157
FcRs Immune complexes ApoJ
Complement activation
C5b-9 C5a
Neutrophils Platelets Macrophages Mesangial cells Cytotoxic
Oxidants Proteases Eicosanoids Cytokines Growth factors Nitric Oxide Epithelial Endothelial Mesangial Proteinuria, GFR, Inflammation, Cell Proliferation, Glomerulosclerosis Proteinuria
Figure 10. Postulated mechanisms of apoJ-mediated suppression of immune
complex-mediated glomerulonephritis.
158 Figure 11. Postulated regulatory role for apoJ in Th1 versus Th2 immune response differentiation. This diagram depicts the differentiation of Th1 versus Th2 immune responses following immune complex-dependent activation. The two immune pathways are distinguished by specific subsets of helper T cells (Th1, Th2), along with their respective cytokines and antibody isotypes. Th1 and Th2 immune responses also have contrasting effects on disease processes. Compared to wild-type mice, apoJ-deficient mice with immune complex-mediated glomerulonephritis develop a polarized Th2 immune response (symbolized by the larger reaction arrows), which correlates with enhanced tissue injury. This observation implies that apoJ regulates the balance between the Th1 and Th2 immune reactions, possibly by modulating immune complex-Fc receptor interactions. ApoJ may also augment the Th1 pathway and/or suppress the Th2 pathway. APC = antigen processing cell. Th0 = undifferentiated helper T cell.
159
Immune complexes IgG2a •Cell-mediated Il-2 immunity
•Delayed ApoJ ? Th1 IFN-γ hypersensitivity
APC Th0 ? ApoJ
Fc receptor •Humoral-mediated Th2 Il-4 immunity
•Autoimmunity Il-10 Enhanced tissue injury ? IgG1
Figure 11. Postulated regulatory role for apoJ in Th1 versus Th2 immune
response differentiation.
160
Final Discussion
The Heat Shock Response and Mouse Apolipoprotein J
161 My studies analyzing the stress inductive and anti-inflammatory properties of the mouse apolipoprotein J (apoJ) gene converge to investigate different facets of the same biological process, namely the heat shock response. The heat shock response is one of the most conserved cellular programs from bacteria to humans (163), and its examination is essential to understanding how an organism reacts to environmental change, tissue injury, and pathogenesis.
The heat shock response not only promotes the survival of individual cells, but emerging evidence has established a tissue protective role for the heat shock response during inflammation, ischemia, and the immune response (163, 240-246).
Hippocrates first described the “beneficial effects of fever” (241). Modern scientists since have observed the protective consequences of the heat shock response in various experimental models. For example, whole body hyperthermia protects rodents against acute respiratory syndrome, myocardial infarction and endotoxin-induced lethality (247-249). Whole organ hyperthermia decreases the harmful effects of ischemia/reperfusion during organ transplantation (250). In addition, heat shock alters leukocyte transmigration across vascular endothelium, thereby limiting neutrophil recruitment and inflammation (243, 249).
Heat shock proteins (HSPs) are the principal mediators of the heat shock response. Via their chaperone activity, HSPs protect cells by binding misfolded and denatured proteins, and dissociating protein aggregates (251). Additionally, HSPs maintain mitochondrial and cellular integrity by inhibiting reactive oxygen species (ROS)-induced lipid peroxidation and DNA strand breaks (240, 241). By limiting the actions of ROS, as well as certain cytokines, HSPs also suppress inflammation.
The individual activities of several HSPs can transcend the protective effects of whole body hyperthermia. Accordingly, cardiac levels of HSP72 are directly associated with
162 myocardial salvage and infarct size after ischemia/reperfusion. In addition, HSP72 over-
expression in transgenic mice protects against myocardial infarction (252). The ectopic expression of another HSP, HSP70, protects against cerebral ischemia and ischemia-like injury
(261). HSPs are also exploited in carcinogenesis to aid in tumor cell survival, and their over-
expression is correlated with de-differentiation, metastasis, and decreased survival rates (253-
257).
The transcription factor, heat shock factor 1 (HSF1), is the master regulator of the heat
shock response in higher eukaryotes (117, 118). It is primed for immediate action in response to
a variety of environmental and pathophysiologic cues, including oxidative stress, fever,
infection, and tissue injury. Like the HSPs, it too has been linked to multiple pathogenic
processes. For example, HSF1 is activated during myocardial and cerebral ischemia (161, 162).
HSF1 is also up-regulated in malignant carcinomas (258). Besides activating HSP gene
transcription, HSF1 represses the expression of several pro-inflammatory cytokines, such as
interleukin-1 (Il-1) and tumor necrosis factor-alpha (TNFα) (259, 260). As a result, disruption of
HSF1-DNA binding activity exacerbates inflammation (244). In addition, HSF1 deficiency
enhances TNFα production, and subsequently increases the lethal effects of endotoxin (122).
Thus, HSF1 and the HSPs are of paramount importance to human disease.
The essence of apoJ function has been mired in ambiguity following many tangential
investigations. Its diverse expression pattern, along with its non-specific protein binding
potential and multiple in vitro functions, have discouraged a clear focus on the physiologic role
of apoJ. However, Easterbrook-Smith and colleagues have embraced the generality of apoJ
biology, and have advanced in the literature an innovative hypothesis pertaining to apoJ function.
They propose that apoJ is a novel secreted HSP based on 3 criteria (32). First, the apoJ promoter
163 contains a highly conserved HSF1-binding motif, i.e. a heat shock element (HSE), which confers stress-induced gene expression (48). Second, the apoJ protein has chaperone-activity, characterized by its affinity for denatured proteins and its ability to prevent protein precipitation
(3). Third, apoJ protects cells from a variety of environmental insults (5, 41).
My thesis strengthens the argument that apoJ is a HSP. I confirmed that the stress
response induction of the mouse apoJ gene requires the apoJ HSE. I also established that HSF1
is essential for apoJ gene activation. In addition, my experiments revealed a cooperative
relationship between the apoJ HSE and an activator protein-1 (AP-1) promoter binding motif,
thereby making the apoJ gene especially responsive to physiologic stress. Consistent with the
anti-inflammatory nature of the heat shock response, I demonstrated that mouse apoJ deficiency
exacerbates immune complex (IC)-mediated glomerulonephritis (GN) in vivo. Thus, in
combination with our myocarditis study (43), my kidney injury work indicates that apoJ is a
general anti-inflammatory factor.
My data, and those of others, establishes apoJ as a prominent player in the heat shock
response. Accordingly, the physiologic role for apoJ is promoted to that of a ubiquitous
protective molecule. Considering the expression profile of apoJ in numerous disease states, apoJ
function should be further examined in experimental models of myocardial infarction,
atherosclerosis, stroke, and autoimmunity. In addition, it would be interesting to investigate possible synergistic and compensating relationships of apoJ with other HSPs. Such relationships could be of biological significance bearing in mind that apoJ-deficient animals develop properly and thrive (Guang Zhu, personal communication). In conclusion, I propose that apoJ is a multifaceted HSP that has evolved in vertebrates to maintain tissue integrity in reaction to inflammation and the immune response.
164 Bibliography
1. Fritz, I. B., Burdzy, K., Setchell, B., and Blaschuk, O. (1983) Ram rete testis fluid
contains a protein (clusterin) which influences cell-cell interactions in vitro. Biol Reprod
28, 1173-1188.
2. Jenne, D. E., and Tschopp, J. (1989) Molecular structure and functional characterization
of a human complement cytolysis inhibitor found in blood and seminal plasma: identity
to sulfated glycoprotein 2, a constituent of rat testis fluid. Proc Natl Acad Sci U S A 86,
7123-7127.
3. Humphreys, D. T., Carver, J. A., Easterbrook-Smith, S. B., and Wilson, M. R. (1999)
Clusterin has chaperone-like activity similar to that of small heat shock proteins. J Biol
Chem 274, 6875-6881.
4. de Silva, H. V., Stuart, W. D., Duvic, C. R., Wetterau, J. R., Ray, M. J., Ferguson, D. G.,
Albers, H. W., Smith, W. R., and Harmony, J. A. (1990) A 70-kDa apolipoprotein
designated ApoJ is a marker for subclasses of human plasma high density lipoproteins. J
Biol Chem 265, 13240-13247.
5. Viard, I., Wehrli, P., Jornot, L., Bullani, R., Vechietti, J. L., Schifferli, J. A., Tschopp, J.,
and French, L. E. (1999) Clusterin gene expression mediates resistance to apoptotic cell
death induced by heat shock and oxidative stress. J Invest Dermatol 112, 290-296.
6. Leger, J. G., Montpetit, M. L., and Tenniswood, M. P. (1987) Characterization and
cloning of androgen-repressed mRNAs from rat ventral prostate. Biochem Biophys Res
Commun 147, 196-203.
165 7. Jordan-Starck, T. C., Witte, D., Aronow, B., and Harmony, J. A. K. (1992)
Apolipoprotein J: a membrane policeman? Current Opinion in Lipidology 3, 75-85.
8. Fritz, I. B., and Burdzy, K. (1989) Novel action of carnitine: inhibition of aggregation of
dispersed cells elicited by clusterin in vitro. J Cell Physiol 140, 18-28.
9. Silkensen, J. R., Skubitz, K. M., Skubitz, A. P., Chmielewski, D. H., Manivel, J. C.,
Dvergsten, J. A., and Rosenberg, M. E. (1995) Clusterin promotes the aggregation and
adhesion of renal porcine epithelial cells. J Clin Invest 96, 2646-2653.
10. Kang, S. W., Lim, S. W., Choi, S. H., Shin, K. H., Chun, B. G., Park, I. S., and Min, B.
H. (2000) Antisense oligonucleotide of clusterin mRNA induces apoptotic cell death and
prevents adhesion of rat ASC-17D Sertoli cells. Mol Cells 10, 193-198.
11. Millis, A. J., Luciani, M., McCue, H. M., Rosenberg, M. E., and Moulson, C. L. (2001)
Clusterin regulates vascular smooth muscle cell nodule formation and migration. J Cell
Physiol 186, 210-219.
12. Gelissen, I. C., Hochgrebe, T., Wilson, M. R., Easterbrook-Smith, S. B., Jessup, W.,
Dean, R. T., and Brown, A. J. (1998) Apolipoprotein J (clusterin) induces cholesterol
export from macrophage- foam cells: a potential anti-atherogenic function? Biochem J
331, 231-237.
13. Van Lenten, B. J., Wagner, A. C., Nayak, D. P., Hama, S., Navab, M., and Fogelman, A.
M. (2001) High-density lipoprotein loses its anti-inflammatory properties during acute
influenza a infection. Circulation 103, 2283-2288.
14. Mackness, B., Hunt, R., Durrington, P. N., and Mackness, M. I. (1997) Increased
immunolocalization of paraoxonase, clusterin, and apolipoprotein A-I in the human artery
166 wall with the progression of atherosclerosis. Arterioscler Thromb Vasc Biol 17, 1233-
1238.
15. Navab, M., Hama-Levy, S., Van Lenten, B. J., Fonarow, G. C., Cardinez, C. J.,
Castellani, L. W., Brennan, M. L., Lusis, A. J., Fogelman, A. M., and La Du, B. N.
(1997) Mildly oxidized LDL induces an increased apolipoprotein J/paraoxonase ratio. J
Clin Invest 99, 2005-2019.
16. Van Lenten, B. J., Wagner, A. C., Navab, M., and Fogelman, A. M. (2001) Oxidized
phospholipids induce changes in hepatic paraoxonase and ApoJ but not monocyte
chemoattractant protein-1 via interleukin-6. J Biol Chem 276, 1923-1929.
17. Bach, U. C., Baiersdorfer, M., Klock, G., Cattaruzza, M., Post, A., and Koch-Brandt, C.
(2001) Apoptotic cell debris and phosphatidylserine-containing lipid vesicles induce
apolipoprotein J (clusterin) gene expression in vital fibroblasts. Exp Cell Res 265, 11-20.
18. Aronow, B. J., Lund, S. D., Brown, T. L., Harmony, J. A., and Witte, D. P. (1993)
Apolipoprotein J expression at fluid-tissue interfaces: potential role in barrier
cytoprotection. Proc Natl Acad Sci U S A 90, 725-729.
19. Murphy, B. F., Kirszbaum, L., Walker, I. D., and d'Apice, A. J. (1988) SP-40,40, a newly
identified normal human serum protein found in the SC5b-9 complex of complement and
in the immune deposits in glomerulonephritis. J Clin Invest 81, 1858-1864.
20. Vakeva, A., Laurila, P., and Meri, S. (1993) Co-deposition of clusterin with the
complement membrane attack complex in myocardial infarction. Immunology 80, 177-
182.
21. Vakeva, A., Meri, S., Lehto, T., and Laurila, P. (1995) Activation of the terminal
complement cascade in renal infarction. Kidney Int 47, 918-926.
167 22. Tschopp, J., Chonn, A., Hertig, S., and French, L. E. (1993) Clusterin, the human
apolipoprotein and complement inhibitor, binds to complement C7, C8 beta, and the b
domain of C9. J Immunol 151, 2159-2165.
23. McDonald, J. F., and Nelsestuen, G. L. (1997) Potent inhibition of terminal complement
assembly by clusterin: characterization of its impact on C9 polymerization. Biochemistry
36, 7464-7473.
24. Saunders, J. R., Aminian, A., McRae, J. L., O'Farrell, K. A., Adam, W. R., and Murphy,
B. F. (1994) Clusterin depletion enhances immune glomerular injury in the isolated
perfused kidney. Kidney Int 45, 817-827.
25. O'Bryan, M. K., Baker, H. W., Saunders, J. R., Kirszbaum, L., Walker, I. D., Hudson, P.,
Liu, D. Y., Glew, M. D., d'Apice, A. J., and Murphy, B. F. (1990) Human seminal
clusterin (SP-40,40). Isolation and characterization. J Clin Invest 85, 1477-1486.
26. Ibrahim, N. M., Troedsson, M. H., Foster, D. N., Loseth, K. J., Farris, J. A., Blaschuk, O.,
and Crabo, B. G. (1999) Reproductive tract secretions and bull spermatozoa contain
different clusterin isoforms that cluster cells and inhibit complement-induced cytolysis. J
Androl 20, 230-240.
27. Sensibar, J. A., Qian, Y., Griswold, M. D., Sylvester, S. R., Bardin, C. W., Cheng, C. Y.,
and Lee, C. (1993) Localization and molecular heterogeneity of sulfated glycoprotein-2
(clusterin) among ventral prostate, seminal vesicle, testis, and epididymis of rats. Biol
Reprod 49, 233-242.
28. O'Bryan, M. K., Murphy, B. F., Liu, D. Y., Clarke, G. N., and Baker, H. W. (1994) The
use of anticlusterin monoclonal antibodies for the combined assessment of human sperm
morphology and acrosome integrity. Hum Reprod 9, 1490-1496.
168 29. Vanderpuye, O. A., Labarrere, C. A., and McIntyre, J. A. (1992) The complement system
in human reproduction. Am J Reprod Immunol 27, 145-155.
30. Ibrahim, N. M., Gilbert, G. R., Loseth, K. J., and Crabo, B. G. (2000) Correlation
between clusterin-positive spermatozoa determined by flow cytometry in bull semen and
fertility. J Androl 21, 887-894.
31. Burkey, B. F., Stuart, W. D., and Harmony, J. A. (1992) Hepatic apolipoprotein J is
secreted as a lipoprotein. J Lipid Res 33, 1517-1526.
32. Wilson, M. R., and Easterbrook-Smith, S. B. (2000) Clusterin is a secreted mammalian
chaperone. Trends Biochem Sci 25, 95-98.
33. Poon, S., Easterbrook-Smith, S. B., Rybchyn, M. S., Carver, J. A., and Wilson, M. R.
(2000) Clusterin is an ATP-independent chaperone with very broad substrate specificity
that stabilizes stressed proteins in a folding-competent state. Biochemistry 39, 15953-
15960.
34. Ghiso, J., Matsubara, E., Koudinov, A., Choi-Miura, N. H., Tomita, M., Wisniewski, T.,
and Frangione, B. (1993) The cerebrospinal-fluid soluble form of Alzheimer's amyloid
beta is complexed to SP-40,40 (apolipoprotein J), an inhibitor of the complement
membrane-attack complex. Biochem J 293, 27-30.
35. Zlokovic, B. V., Martel, C. L., Matsubara, E., McComb, J. G., Zheng, G., McCluskey, R.
T., Frangione, B., and Ghiso, J. (1996) Glycoprotein 330/megalin: probable role in
receptor-mediated transport of apolipoprotein J alone and in a complex with Alzheimer
disease amyloid beta at the blood-brain and blood-cerebrospinal fluid barriers. Proc Natl
Acad Sci U S A 93, 4229-4234.
169 36. Matsubara, E., Soto, C., Governale, S., Frangione, B., and Ghiso, J. (1996)
Apolipoprotein J and Alzheimer's amyloid beta solubility. Biochem J 316, 671-679.
37. Choi-Miura, N. H., and Oda, T. (1996) Relationship between multifunctional protein
"clusterin" and Alzheimer disease. Neurobiol Aging 17, 717-722.
38. Boggs, L. N., Fuson, K. S., Baez, M., Churgay, L., McClure, D., Becker, G., and May, P.
C. (1996) Clusterin (Apo J) protects against in vitro amyloid-beta (1-40) neurotoxicity. J
Neurochem 67, 1324-1327.
39. Giannakopoulos, P., Kovari, E., French, L. E., Viard, I., Hof, P. R., and Bouras, C. (1998)
Possible neuroprotective role of clusterin in Alzheimer's disease: a quantitative
immunocytochemical study. Acta Neuropathol (Berl) 95, 387-394.
40. Lidstrom, A. M., Bogdanovic, N., Hesse, C., Volkman, I., Davidsson, P., and Blennow,
K. (1998) Clusterin (apolipoprotein J) protein levels are increased in hippocampus and in
frontal cortex in Alzheimer's disease. Exp Neurol 154, 511-521.
41. Miyake, H., Nelson, C., Rennie, P. S., and Gleave, M. E. (2000) Testosterone-repressed
prostate message-2 is an antiapoptotic gene involved in progression to androgen
independence in prostate cancer. Cancer Res 60, 170-176.
42. Newkirk, M. M., Apostolakos, P., Neville, C., and Fortin, P. R. (1999) Systemic lupus
erythematosus, a disease associated with low levels of clusterin/apoJ, an
antiinflammatory protein. J Rheumatol 26, 597-603.
43. McLaughlin, L., Zhu, G., Mistry, M., Ley-Ebert, C., Stuart, W. D., Florio, C. J., Groen,
P. A., Witt, S. A., Kimball, T. R., Witte, D. P., Harmony, J. A., and Aronow, B. J. (2000)
Apolipoprotein J/clusterin limits the severity of murine autoimmune myocarditis. J Clin
Invest 106, 1105-1113.
170 44. Jordan-Starck, T. C., Lund, S. D., Witte, D. P., Aronow, B. J., Ley, C. A., Stuart, W. D.,
Swertfeger, D. K., Clayton, L. R., Sells, S. F., Paigen, B., and et al. (1994) Mouse
apolipoprotein J: characterization of a gene implicated in atherosclerosis. J Lipid Res 35,
194-210.
45. Fink, T. M., Zimmer, M., Tschopp, J., Etienne, J., Jenne, D. E., and Lichter, P. (1993)
Human clusterin (CLI) maps to 8p21 in proximity to the lipoprotein lipase (LPL) gene.
Genomics 16, 526-528.
46. Nestlerode, C. S., Bunker, C. H., Sanghera, D. K., Aston, C. E., Ukoli, F. A., and
Kamboh, M. I. (1999) Apolipoprotein J polymorphisms and serum HDL cholesterol
levels in African blacks. Hum Biol 71, 197-218.
47. Tycko, B., Feng, L., Nguyen, L., Francis, A., Hays, A., Chung, W. Y., Tang, M. X.,
Stern, Y., Sahota, A., Hendrie, H., and Mayeux, R. (1996) Polymorphisms in the human
apolipoprotein-J/clusterin gene: ethnic variation and distribution in Alzheimer's disease.
Hum Genet 98, 430-436.
48. Michel, D., Chatelain, G., North, S., and Brun, G. (1997) Stress-induced transcription of
the clusterin/apoJ gene. Biochem J 328, 45-50.
49. de Silva, H. V., Harmony, J. A., Stuart, W. D., Gil, C. M., and Robbins, J. (1990)
Apolipoprotein J: structure and tissue distribution. Biochemistry 29, 5380-5389.
50. Rosenberg, M. E., and Silkensen, J. (1995) Clusterin: physiologic and pathophysiologic
considerations. Int J Biochem Cell Biol 27, 633-645.
51. Reddy, K. B., Jin, G., Karode, M. C., Harmony, J. A., and Howe, P. H. (1996)
Transforming growth factor beta (TGF beta)-induced nuclear localization of
apolipoprotein J/clusterin in epithelial cells. Biochemistry 35, 6157-6163.
171 52. Yang, C. R., Leskov, K., Hosley-Eberlein, K., Criswell, T., Pink, J. J., Kinsella, T. J., and
Boothman, D. A. (2000) Nuclear clusterin/XIP8, an x-ray-induced Ku70-binding protein
that signals cell death. Proc Natl Acad Sci U S A 97, 5907-5912.
53. Jenne, D. E., Lowin, B., Peitsch, M. C., Bottcher, A., Schmitz, G., and Tschopp, J. (1991)
Clusterin (complement lysis inhibitor) forms a high density lipoprotein complex with
apolipoprotein A-I in human plasma. J Biol Chem 266, 11030-11036.
54. Stuart, W. D., Krol, B., Jenkins, S. H., and Harmony, J. A. (1992) Structure and stability
of apolipoprotein J-containing high-density lipoproteins. Biochemistry 31, 8552-8559.
55. Shanmugaratnam, J., Berg, E., Kimerer, L., Johnson, R. J., Amaratunga, A., Schreiber, B.
M., and Fine, R. E. (1997) Retinal Muller glia secrete apolipoproteins E and J which are
efficiently assembled into lipoprotein particles. Brain Res Mol Brain Res 50, 113-120.
56. Xu, Q., Li, Y., Cyras, C., Sanan, D. A., and Cordell, B. (2000) Isolation and
characterization of apolipoproteins from murine microglia. Identification of a low density
lipoprotein-like apolipoprotein J-rich but E-poor spherical particle. J Biol Chem 275,
31770-31777.
57. O'Bryan, M. K., Mallidis, C., Murphy, B. F., and Baker, H. W. (1994)
Immunohistological localization of clusterin in the male genital tract in humans and
marmosets. Biol Reprod 50, 502-509.
58. Harding, M. A., Chadwick, L. J., Gattone, V. H., 2nd, and Calvet, J. P. (1991) The SGP-2
gene is developmentally regulated in the mouse kidney and abnormally expressed in
collecting duct cysts in polycystic kidney disease. Dev Biol 146, 483-490.
59. Reeder, D. J., Stuart, W. D., Witte, D. P., Brown, T. L., and Harmony, J. A. (1995) Local
synthesis of apolipoprotein J in the eye. Exp Eye Res 60, 495-504.
172 60. Witte, D. P., Aronow, B. J., Stauderman, M. L., Stuart, W. D., Clay, M. A., Gruppo, R.
A., Jenkins, S. H., and Harmony, J. A. (1993) Platelet activation releases megakaryocyte-
synthesized apolipoprotein J, a highly abundant protein in atheromatous lesions. Am J
Pathol 143, 763-773.
61. Buttyan, R., Olsson, C. A., Pintar, J., Chang, C., Bandyk, M., Ng, P. Y., and Sawczuk, I.
S. (1989) Induction of the TRPM-2 gene in cells undergoing programmed death. Mol
Cell Biol 9, 3473-3481.
62. Bandyk, M. G., Sawczuk, I. S., Olsson, C. A., Katz, A. E., and Buttyan, R. (1990)
Characterization of the products of a gene expressed during androgen- programmed cell
death and their potential use as a marker of urogenital injury. J Urol 143, 407-413.
63. Tenniswood, M. P., Guenette, R. S., Lakins, J., Mooibroek, M., Wong, P., and Welsh, J.
E. (1992) Active cell death in hormone-dependent tissues. Cancer Metastasis Rev 11,
197-220.
64. Guenette, R. S., Corbeil, H. B., Leger, J., Wong, K., Mezl, V., Mooibroek, M., and
Tenniswood, M. (1994) Induction of gene expression during involution of the lactating
mammary gland of the rat. J Mol Endocrinol 12, 47-60.
65. French, L. E., Soriano, J. V., Montesano, R., and Pepper, M. S. (1996) Modulation of
clusterin gene expression in the rat mammary gland during pregnancy, lactation, and
involution. Biol Reprod 55, 1213-1220.
66. Brown, T. L., Moulton, B. C., Witte, D. P., Swertfeger, D. K., and Harmony, J. A. (1996)
Apolipoprotein J/clusterin expression defines distinct stages of blastocyst implantation in
the mouse uterus. Biol Reprod 55, 740-747.
173 67. French, L. E., Chonn, A., Ducrest, D., Baumann, B., Belin, D., Wohlwend, A., Kiss, J. Z.,
Sappino, A. P., Tschopp, J., and Schifferli, J. A. (1993) Murine clusterin: molecular
cloning and mRNA localization of a gene associated with epithelial differentiation
processes during embryogenesis. J Cell Biol 122, 1119-1130.
68. Witte, D. P., Aronow, B. J., Dry, J. K., and Harmony, J. A. (1994) Temporally and
spatially restricted expression of apolipoprotein J in the developing heart defines discrete
stages of valve morphogenesis. Dev Dyn 201, 290-296.
69. Clark, A. M., and Griswold, M. D. (1997) Expression of clusterin/sulfated glycoprotein-2
under conditions of heat stress in rat Sertoli cells and a mouse Sertoli cell line. J Androl
18, 257-263.
70. Rozovsky, I., Morgan, T. E., Willoughby, D. A., Dugichi-Djordjevich, M. M., Pasinetti,
G. M., Johnson, S. A., and Finch, C. E. (1994) Selective expression of clusterin (SGP-2)
and complement C1qB and C4 during responses to neurotoxins in vivo and in vitro.
Neuroscience 62, 741-758.
71. Pasinetti, G. M., Cheng, H. W., Morgan, D. G., Lampert-Etchells, M., McNeill, T. H.,
and Finch, C. E. (1993) Astrocytic messenger RNA responses to striatal deafferentation
in male rat. Neuroscience 53, 199-211.
72. Han, B. H., DeMattos, R. B., Dugan, L. L., Kim-Han, J. S., Brendza, R. P., Fryer, J. D.,
Kierson, M., Cirrito, J., Quick, K., Harmony, J. A., Aronow, B. J., and Holtzman, D. M.
(2001) Clusterin contributes to caspase-3-independent brain injury following neonatal
hypoxia-ischemia. Nat Med 7, 338-343.
174 73. Van Beek, J., Chan, P., Bernaudin, M., Petit, E., MacKenzie, E. T., and Fontaine, M.
(2000) Glial responses, clusterin, and complement in permanent focal cerebral ischemia
in the mouse. Glia 31, 39-50.
74. Walton, M., Young, D., Sirimanne, E., Dodd, J., Christie, D., Williams, C., Gluckman,
P., and Dragunow, M. (1996) Induction of clusterin in the immature brain following a
hypoxic- ischemic injury. Brain Res Mol Brain Res 39, 137-152.
75. Yamashita, K., Vogel, P., Fritze, K., Back, T., Hossmann, K. A., and Wiessner, C. (1996)
Monitoring the temporal and spatial activation pattern of astrocytes in focal cerebral
ischemia using in situ hybridization to GFAP mRNA: comparison with sgp-2 and hsp70
mRNA and the effect of glutamate receptor antagonists. Brain Res 735, 285-297.
76. Yasuhara, O., Aimi, Y., Yamada, T., Matsuo, A., McGeer, E. G., and McGeer, P. L.
(1994) Clusterin as a marker for ischaemic Purkinje cells in human brain.
Neurodegeneration 3, 325-329.
77. May, P. C., and Finch, C. E. (1992) Sulfated glycoprotein 2: new relationships of this
multifunctional protein to neurodegeneration. Trends Neurosci 15, 391-396.
78. Grewal, R. P., Morgan, T. E., and Finch, C. E. (1999) C1qB and clusterin mRNA
increase in association with neurodegeneration in sporadic amyotrophic lateral sclerosis.
Neurosci Lett 271, 65-67.
79. Duguid, J. R., Bohmont, C. W., Liu, N. G., and Tourtellotte, W. W. (1989) Changes in
brain gene expression shared by scrapie and Alzheimer disease. Proc Natl Acad Sci U S A
86, 7260-7264.
175 80. May, P. C., Lampert-Etchells, M., Johnson, S. A., Poirier, J., Masters, J. N., and Finch, C.
E. (1990) Dynamics of gene expression for a hippocampal glycoprotein elevated in
Alzheimer's disease and in response to experimental lesions in rat. Neuron 5, 831-839.
81. Kida, E., Choi-Miura, N. H., and Wisniewski, K. E. (1995) Deposition of apolipoproteins
E and J in senile plaques is topographically determined in both Alzheimer's disease and
Down's syndrome brain. Brain Res 685, 211-216.
82. Ishikawa, Y., Akasaka, Y., Ishii, T., Komiyama, K., Masuda, S., Asuwa, N., Choi-Miura,
N. H., and Tomita, M. (1998) Distribution and synthesis of apolipoprotein J in the
atherosclerotic aorta. Arterioscler Thromb Vasc Biol 18, 665-672.
83. Silkensen, J. R., Hirsch, A. T., Lunzer, M. M., Chmielewski, D., Manivel, J. C.,
Muellerleile, M. R., and Rosenberg, M. E. (1998) Temporal induction of clusterin in the
peri-infarct zone after experimental myocardial infarction in the rat. J Lab Clin Med 131,
28-35.
84. Swertfeger, D. K., Witte, D. P., Stuart, W. D., Rockman, H. A., and Harmony, J. A.
(1996) Apolipoprotein J/clusterin induction in myocarditis: A localized response gene to
myocardial injury. Am J Pathol 148, 1971-1983.
85. Aulitzky, W. K., Schlegel, P. N., Wu, D. F., Cheng, C. Y., Chen, C. L., Li, P. S.,
Goldstein, M., Reidenberg, M., and Bardin, C. W. (1992) Measurement of urinary
clusterin as an index of nephrotoxicity. Proc Soc Exp Biol Med 199, 93-96.
86. Silkensen, J. R., Agarwal, A., Nath, K. A., Manivel, J. C., and Rosenberg, M. E. (1997)
Temporal induction of clusterin in cisplatin nephrotoxicity. J Am Soc Nephrol 8, 302-
305.
176 87. Correa-Rotter, R., Hostetter, T. H., Nath, K. A., Manivel, J. C., and Rosenberg, M. E.
(1992) Interaction of complement and clusterin in renal injury. J Am Soc Nephrol 3,
1172-1179.
88. Rosenberg, M. E., and Paller, M. S. (1991) Differential gene expression in the recovery
from ischemic renal injury. Kidney Int 39, 1156-1161.
89. Dvergsten, J., Manivel, J. C., Correa-Rotter, R., and Rosenberg, M. E. (1994) Expression
of clusterin in human renal diseases. Kidney Int 45, 828-835.
90. Gattone, V. H., 2nd, Lowden, D. A., and Cowley, B. D., Jr. (1995) Epidermal growth
factor ameliorates autosomal recessive polycystic kidney disease in mice. Dev Biol 169,
504-510.
91. Rosenberg, M. E., Manivel, J. C., Carone, F. A., and Kanwar, Y. S. (1995) Genesis of
renal cysts is associated with clusterin expression in experimental cystic disease. J Am
Soc Nephrol 5, 1669-1674.
92. Ricker, J. L., Gattone, V. H., 2nd, Calvet, J. P., and Rankin, C. A. (2000) Development of
autosomal recessive polycystic kidney disease in BALB/c- cpk/cpk mice. J Am Soc
Nephrol 11, 1837-1847.
93. Murphy, B. F., Davies, D. J., Morrow, W., and d'Apice, A. J. (1989) Localization of
terminal complement components S-protein and SP-40,40 in renal biopsies. Pathology
21, 275-278.
94. French, L. E., Tschopp, J., and Schifferli, J. A. (1992) Clusterin in renal tissue:
preferential localization with the terminal complement complex and immunoglobulin
deposits in glomeruli. Clin Exp Immunol 88, 389-393.
177 95. Kadomatsu, K., Anzano, M. A., Slayter, M. V., Winokur, T. S., Smith, J. M., and Sporn,
M. B. (1993) Expression of sulfated glycoprotein 2 is associated with carcinogenesis
induced by N-nitroso-N-methylurea in rat prostate and seminal vesicle. Cancer Res 53,
1480-1483.
96. Parczyk, K., Pilarsky, C., Rachel, U., and Koch-Brandt, C. (1994) Gp80 (clusterin;
TRPM-2) mRNA level is enhanced in human renal clear cell carcinomas. J Cancer Res
Clin Oncol 120, 186-188.
97. Steinberg, J., Oyasu, R., Lang, S., Sintich, S., Rademaker, A., Lee, C., Kozlowski, J. M.,
and Sensibar, J. A. (1997) Intracellular levels of SGP-2 (Clusterin) correlate with tumor
grade in prostate cancer. Clin Cancer Res 3, 1707-1711.
98. Wellmann, A., Thieblemont, C., Pittaluga, S., Sakai, A., Jaffe, E. S., Siebert, P., and
Raffeld, M. (2000) Detection of differentially expressed genes in lymphomas using
cDNA arrays: identification of clusterin as a new diagnostic marker for anaplastic large-
cell lymphomas. Blood 96, 398-404.
99. Hough, C. D., Cho, K. R., Zonderman, A. B., Schwartz, D. R., and Morin, P. J. (2001)
Coordinately up-regulated genes in ovarian cancer. Cancer Res 61, 3869-3876.
100. Danik, M., Chabot, J. G., Mercier, C., Benabid, A. L., Chauvin, C., Quirion, R., and Suh,
M. (1991) Human gliomas and epileptic foci express high levels of a mRNA related to rat
testicular sulfated glycoprotein 2, a purported marker of cell death. Proc Natl Acad Sci U
S A 88, 8577-8581.
101. Redondo, M., Villar, E., Torres-Munoz, J., Tellez, T., Morell, M., and Petito, C. K.
(2000) Overexpression of clusterin in human breast carcinoma. Am J Pathol 157, 393-
399.
178 102. Jin, G., and Howe, P. H. (1999) Transforming growth factor beta regulates clusterin gene
expression via modulation of transcription factor c-Fos. Eur J Biochem 263, 534-542.
103. Jin, G., and Howe, P. H. (1997) Regulation of clusterin gene expression by transforming
growth factor beta. J Biol Chem 272, 26620-26626.
104. Gutacker, C., Klock, G., Diel, P., and Koch-Brandt, C. (1999) Nerve growth factor and
epidermal growth factor stimulate clusterin gene expression in PC12 cells. Biochem J
339, 759-766.
105. Cervellera, M., Raschella, G., Santilli, G., Tanno, B., Ventura, A., Mancini, C.,
Sevignani, C., Calabretta, B., and Sala, A. (2000) Direct transactivation of the anti-
apoptotic gene apolipoprotein J (clusterin) by B-MYB. J Biol Chem 275, 21055-21060.
106. Furlong, E. E., Keon, N. K., Thornton, F. D., Rein, T., and Martin, F. (1996) Expression
of a 74-kDa nuclear factor 1 (NF1) protein is induced in mouse mammary gland
involution. Involution-enhanced occupation of a twin NF1 binding element in the
testosterone-repressed prostate message- 2/clusterin promoter. J Biol Chem 271, 29688-
29697.
107. Lymar, E. S., Clark, A. M., Reeves, R., and Griswold, M. D. (2000) Clusterin gene in rat
sertoli cells is regulated by a core-enhancer element. Biol Reprod 63, 1341-1351.
108. Michel, D., Chatelain, G., Herault, Y., and Brun, G. (1995) The expression of the avian
clusterin gene can be driven by two alternative promoters with distinct regulatory
elements. Eur J Biochem 229, 215-223.
109. Johnson, A. D. (1995) The price of repression. Cell 81, 655-658.
179 110. Calero, M., Rostagno, A., Matsubara, E., Zlokovic, B., Frangione, B., and Ghiso, J.
(2000) Apolipoprotein J (clusterin) and Alzheimer's disease. Microsc Res Tech 50, 305-
315.
111. Hochgrebe, T., Pankhurst, G. J., Wilce, J., and Easterbrook-Smith, S. B. (2000) pH-
dependent changes in the in vitro ligand-binding properties and structure of human
clusterin. Biochemistry 39, 1411-1419.
112. Ahuja, H. S., Tenniswood, M., Lockshin, R., and Zakeri, Z. F. (1994) Expression of
clusterin in cell differentiation and cell death. Biochem Cell Biol 72, 523-530.
113. Little, S. A., and Mirkes, P. E. (1995) Clusterin expression during programmed and
teratogen-induced cell death in the postimplantation rat embryo. Teratology 52, 41-54.
114. Min, B. H., Jeong, S. Y., Kang, S. W., Crabo, B. G., Foster, D. N., Chun, B. G.,
Bendayan, M., and Park, I. S. (1998) Transient expression of clusterin (sulfated
glycoprotein-2) during development of rat pancreas. J Endocrinol 158, 43-52.
115. O'Bryan, M. K., Cheema, S. S., Bartlett, P. F., Murphy, B. F., and Pearse, M. J. (1993)
Clusterin levels increase during neuronal development. J Neurobiol 24, 421-432.
116. Moll, S., Menoud, P. A., French, L., Sappino, A. P., Pastore, Y., Schifferli, J. A., and
Izui, S. (1998) Tubular up-regulation of clusterin mRNA in murine lupus-like nephritis.
Am J Pathol 152, 953-962.
117. Pirkkala, L., Nykanen, P., and Sistonen, L. (2001) Roles of the heat shock transcription
factors in regulation of the heat shock response and beyond. Faseb J 15, 1118-1131.
118. Morimoto, R. I. (1998) Regulation of the heat shock transcriptional response: cross talk
between a family of heat shock factors, molecular chaperones, and negative regulators.
Genes Dev 12, 3788-3796.
180 119. Wilson, M. R., and Easterbrook-Smith, S. B. (1992) Clusterin binds by a multivalent
mechanism to the Fc and Fab regions of IgG. Biochim Biophys Acta 1159, 319-326.
120. Wilson, M. R., Roeth, P. J., and Easterbrook-Smith, S. B. (1991) Clusterin enhances the
formation of insoluble immune complexes. Biochem Biophys Res Commun 177, 985-990.
121. Pankhurst, G. J., Bennett, C. A., and Easterbrook-Smith, S. B. (1998) Characterization of
the heparin-binding properties of human clusterin. Biochemistry 37, 4823-4830.
122. Xiao, X., Zuo, X., Davis, A. A., McMillan, D. R., Curry, B. B., Richardson, J. A., and
Benjamin, I. J. (1999) HSF1 is required for extra-embryonic development, postnatal
growth and protection during inflammatory responses in mice. Embo J 18, 5943-5952.
123. McMillan, D. R., Xiao, X., Shao, L., Graves, K., and Benjamin, I. J. (1998) Targeted
disruption of heat shock transcription factor 1 abolishes thermotolerance and protection
against heat-inducible apoptosis. J Biol Chem 273, 7523-7528.
124. Kroeger, P. E., and Morimoto, R. I. (1994) Selection of new HSF1 and HSF2 DNA-
binding sites reveals difference in trimer cooperativity. Mol Cell Biol 14, 7592-7603.
125. Kroeger, P. E., Sarge, K. D., and Morimoto, R. I. (1993) Mouse heat shock transcription
factors 1 and 2 prefer a trimeric binding site but interact differently with the HSP70 heat
shock element. Mol Cell Biol 13, 3370-3383.
126. Bernelli-Zazzera, A., Cairo, G., Schiaffonati, L., and Tacchini, L. (1992) Stress proteins
and reperfusion stress in the liver. Ann N Y Acad Sci 663, 120-124.
127. Fabbro, C., Braghetta, P., Girotto, D., Piccolo, S., Volpin, D., and Bressan, G. M. (1999)
Cell type-specific transcription of the alpha1(VI) collagen gene. Role of the AP1 binding
site and of the core promoter. J Biol Chem 274, 1759-1766.
181 128. Quinones, S., Buttice, G., and Kurkinen, M. (1994) Promoter elements in the
transcriptional activation of the human stromelysin-1 gene by the inflammatory cytokine,
interleukin 1. Biochem J 302, 471-477.
129. Erdos, G., Lee, Y. J., Cho, J. M., and Corry, P. M. (1995) Heat-induced bFGF gene
expression in the absence of heat shock element correlates with enhanced AP-1 binding
activity. J Cell Physiol 164, 404-413.
130. Shi, Q., Le, X., Abbruzzese, J. L., Wang, B., Mujaida, N., Matsushima, K., Huang, S.,
Xiong, Q., and Xie, K. (1999) Cooperation between transcription factor AP-1 and NF-
kappaB in the induction of interleukin-8 in human pancreatic adenocarcinoma cells by
hypoxia. J Interferon Cytokine Res 19, 1363-1371.
131. Diamond, D. A., Parsian, A., Hunt, C. R., Lofgren, S., Spitz, D. R., Goswami, P. C., and
Gius, D. (1999) Redox factor-1 (Ref-1) mediates the activation of AP-1 in HeLa and NIH
3T3 cells in response to heat shock. J Biol Chem 274, 16959-16964.
132. Adler, V., Schaffer, A., Kim, J., Dolan, L., and Ronai, Z. (1995) UV irradiation and heat
shock mediate JNK activation via alternate pathways. J Biol Chem 270, 26071-26077.
133. Boshart, M., Kluppel, M., Schmidt, A., Schutz, G., and Luckow, B. (1992) Reporter
constructs with low background activity utilizing the cat gene. Gene 110, 129-130.
134. Rabindran, S. K., Giorgi, G., Clos, J., and Wu, C. (1991) Molecular cloning and
expression of a human heat shock factor, HSF1. Proc Natl Acad Sci U S A 88, 6906-
6910.
135. Aronow, B. J., Silbiger, R. N., Dusing, M. R., Stock, J. L., Yager, K. L., Potter, S. S.,
Hutton, J. J., and Wiginton, D. A. (1992) Functional analysis of the human adenosine
182 deaminase gene thymic regulatory region and its ability to generate position-independent
transgene expression. Mol Cell Biol 12, 4170-4185.
136. Hogan, B. (1983) Molecular biology. Enhancers, chromosome position effects, and
transgenic mice [news]. Nature 306, 313-314.
137. Sambrook, J., Fritsch, E. F., and Maniatis, T., eds (1989) Molecular Cloning, A
Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New
York.
138. Schwartz, S., Zhang, Z., Frazer, K. A., Smit, A., Riemer, C., Bouck, J., Gibbs, R.,
Hardison, R., and Miller, W. (2000) PipMaker--a web server for aligning two genomic
DNA sequences. Genome Res 10, 577-586.
139. Smit, A. F. A., and Green, P. (1999) Repeat Masker Website. In
140. Quandt, K., Frech, K., Karas, H., Wingender, E., and Werner, T. (1995) MatInd and
MatInspector: new fast and versatile tools for detection of consensus matches in
nucleotide sequence data. Nucleic Acids Res 23, 4878-4884.
141. Wingender, E., Chen, X., Hehl, R., Karas, H., Liebich, I., Matys, V., Meinhardt, T.,
Pruss, M., Reuter, I., and Schacherer, F. (2000) TRANSFAC: an integrated system for
gene expression regulation. Nucleic Acids Res 28, 316-319.
142. Chu, B., Zhong, R., Soncin, F., Stevenson, M. A., and Calderwood, S. K. (1998)
Transcriptional activity of heat shock factor 1 at 37 degrees C is repressed through
phosphorylation on two distinct serine residues by glycogen synthase kinase 3 and
protein kinases Calpha and Czeta. J Biol Chem 273, 18640-18646.
183 143. Abravaya, K., Phillips, B., and Morimoto, R. I. (1991) Heat shock-induced interactions of
heat shock transcription factor and the human hsp70 promoter examined by in vivo
footprinting. Mol Cell Biol 11, 586-592.
144. Newton, E. M., Knauf, U., Green, M., and Kingston, R. E. (1996) The regulatory domain
of human heat shock factor 1 is sufficient to sense heat stress. Mol Cell Biol 16, 839-846.
145. Xiao, H., Perisic, O., and Lis, J. T. (1991) Cooperative binding of Drosophila heat shock
factor to arrays of a conserved 5 bp unit. Cell 64, 585-593.
146. Pitts, R. L., Wang, S., Jones, E. A., and Symes, A. J. (2001) Transforming growth factor-
beta and ciliary neurotrophic factor synergistically induce vasoactive intestinal peptide
gene expression through the cooperation of Smad, STAT, and AP-1 sites. J Biol Chem
276, 19966-19973.
147. Bergelson, S., and Daniel, V. (1994) Cooperative interaction between Ets and AP-1
transcription factors regulates induction of glutathione S-transferase Ya gene expression.
Biochem Biophys Res Commun 200, 290-297.
148. D'Orazio, D., Besser, D., Marksitzer, R., Kunz, C., Hume, D. A., Kiefer, B., and
Nagamine, Y. (1997) Cooperation of two PEA3/AP1 sites in uPA gene induction by TPA
and FGF- 2. Gene 201, 179-187.
149. Ng, K. W., Ridgway, P., Cohen, D. R., and Tremethick, D. J. (1997) The binding of a
Fos/Jun heterodimer can completely disrupt the structure of a nucleosome. Embo J 16,
2072-2085.
150. Ito, T., Yamauchi, M., Nishina, M., Yamamichi, N., Mizutani, T., Ui, M., Murakami, M.,
and Iba, H. (2001) Identification of SWI.SNF complex subunit BAF60a as a determinant
of the transactivation potential of Fos/Jun dimers. J Biol Chem 276, 2852-2857.
184 151. Rosemblit, N., and Chen, C. L. (1994) Regulators for the rat clusterin gene: DNA
methylation and cis-acting regulatory elements. J Mol Endocrinol 13, 69-76.
152. Susek, R. E., and Lindquist, S. (1990) Transcriptional derepression of the Saccharomyces
cerevisiae HSP26 gene during heat shock. Mol Cell Biol 10, 6362-6373.
153. Wunsche, W., Tenniswood, M. P., Schneider, M. R., and Vollmer, G. (1998) Estrogenic
regulation of clusterin mRNA in normal and malignant endometrial tissue. Int J Cancer
76, 684-688.
154. Ogbourne, S., and Antalis, T. M. (1998) Transcriptional control and the role of silencers
in transcriptional regulation in eukaryotes. Biochem J 331, 1-14.
155. Smith, C. L., and Hager, G. L. (1997) Transcriptional regulation of mammalian genes in
vivo. A tale of two templates. J Biol Chem 272, 27493-27496.
156. Chen, G., and Courey, A. J. (2000) Groucho/TLE family proteins and transcriptional
repression. Gene 249, 1-16.
157. Tolkunova, E. N., Fujioka, M., Kobayashi, M., Deka, D., and Jaynes, J. B. (1998) Two
distinct types of repression domain in engrailed: one interacts with the groucho
corepressor and is preferentially active on integrated target genes. Mol Cell Biol 18,
2804-2814.
158. Lekstrom-Himes, J., and Xanthopoulos, K. G. (1998) Biological role of the
CCAAT/enhancer-binding protein family of transcription factors. J Biol Chem 273,
28545-28548.
159. Poli, V. (1998) The role of C/EBP isoforms in the control of inflammatory and native
immunity functions. J Biol Chem 273, 29279-29282.
185 160. Sabatakos, G., Davies, G. E., Grosse, M., Cryer, A., and Ramji, D. P. (1998) Expression
of the genes encoding CCAAT-enhancer binding protein isoforms in the mouse
mammary gland during lactation and involution. Biochem J 334, 205-210.
161. Nishizawa, J., Nakai, A., Matsuda, K., Komeda, M., Ban, T., and Nagata, K. (1999)
Reactive oxygen species play an important role in the activation of heat shock factor 1 in
ischemic-reperfused heart. Circulation 99, 934-941.
162. Higashi, T., Nakai, A., Uemura, Y., Kikuchi, H., and Nagata, K. (1995) Activation of
heat shock factor 1 in rat brain during cerebral ischemia or after heat shock. Brain Res
Mol Brain Res 34, 262-270.
163. Williams, R. S., and Benjamin, I. J. (2000) Protective responses in the ischemic
myocardium. J Clin Invest 106, 813-818.
164. Jaggi, R., Marti, A., Guo, K., Feng, Z., and Friis, R. R. (1996) Regulation of a
physiological apoptosis: mouse mammary involution. J Dairy Sci 79, 1074-1084.
165. Marti, A., Jehn, B., Costello, E., Keon, N., Ke, G., Martin, F., and Jaggi, R. (1994)
Protein kinase A and AP-1 (c-Fos/JunD) are induced during apoptosis of mouse
mammary epithelial cells. Oncogene 9, 1213-1223.
166. Marti, A., Lazar, H., Ritter, P., and Jaggi, R. (1999) Transcription factor activities and
gene expression during mouse mammary gland involution. J Mammary Gland Biol
Neoplasia 4, 145-152.
167. Eriksson, M., Jokinen, E., Sistonen, L., and Leppa, S. (2000) Heat shock factor 2 is
activated during mouse heart development. Int J Dev Biol 44, 471-477.
186 168. Goodson, M. L., Park-Sarge, O. K., and Sarge, K. D. (1995) Tissue-dependent expression
of heat shock factor 2 isoforms with distinct transcriptional activities. Mol Cell Biol 15,
5288-5293.
169. Loones, M. T., Rallu, M., Mezger, V., and Morange, M. (1997) HSP gene expression and
HSF2 in mouse development. Cell Mol Life Sci 53, 179-190.
170. Rallu, M., Loones, M., Lallemand, Y., Morimoto, R., Morange, M., and Mezger, V.
(1997) Function and regulation of heat shock factor 2 during mouse embryogenesis. Proc
Natl Acad Sci U S A 94, 2392-2397.
171. Sarge, K. D., Park-Sarge, O. K., Kirby, J. D., Mayo, K. E., and Morimoto, R. I. (1994)
Expression of heat shock factor 2 in mouse testis: potential role as a regulator of heat-
shock protein gene expression during spermatogenesis. Biol Reprod 50, 1334-1343.
172. Aronow, B., Lattier, D., Silbiger, R., Dusing, M., Hutton, J., Jones, G., Stock, J.,
McNeish, J., Potter, S., Witte, D., and et al. (1989) Evidence for a complex regulatory
array in the first intron of the human adenosine deaminase gene. Genes Dev 3, 1384-1400
173. Talbot, D., and Grosveld, F. (1991) The 5'HS2 of the globin locus control region
enhances transcription through the interaction of a multimeric complex binding at two
functionally distinct NF-E2 binding sites. Embo J 10, 1391-1398.
174. Kim, C. G., Epner, E. M., Forrester, W. C., and Groudine, M. (1992) Inactivation of the
human beta-globin gene by targeted insertion into the beta-globin locus control region.
Genes Dev 6, 928-938.
175. Lander, E. S., Linton, L. M., Birren, B., Nusbaum, C., Zody, M. C., Baldwin, J., Devon,
K., Dewar, K., Doyle, M., FitzHugh, W., Funke, R., Gage, D., Harris, K., Heaford, A.,
Howland, J., Kann, L., Lehoczky, J., LeVine, R., McEwan, P., McKernan, K., Meldrim,
187 J., Mesirov, J. P., Miranda, C., Morris, W., Naylor, J., Raymond, C., Rosetti, M., Santos,
R., Sheridan, A., Sougnez, C., Stange-Thomann, N., Stojanovic, N., Subramanian, A.,
Wyman, D., Rogers, J., Sulston, J., Ainscough, R., Beck, S., Bentley, D., Burton, J., Clee,
C., Carter, N., Coulson, A., Deadman, R., Deloukas, P., Dunham, A., Dunham, I.,
Durbin, R., French, L., Grafham, D., Gregory, S., Hubbard, T., Humphray, S., Hunt, A.,
Jones, M., Lloyd, C., McMurray, A., Matthews, L., Mercer, S., Milne, S., Mullikin, J. C.,
Mungall, A., Plumb, R., Ross, M., Shownkeen, R., Sims, S., Waterston, R. H., Wilson, R.
K., Hillier, L. W., McPherson, J. D., Marra, M. A., Mardis, E. R., Fulton, L. A.,
Chinwalla, A. T., Pepin, K. H., Gish, W. R., Chissoe, S. L., Wendl, M. C., Delehaunty,
K. D., Miner, T. L., Delehaunty, A., Kramer, J. B., Cook, L. L., Fulton, R. S., Johnson,
D. L., Minx, P. J., Clifton, S. W., Hawkins, T., Branscomb, E., Predki, P., Richardson, P.,
Wenning, S., Slezak, T., Doggett, N., Cheng, J. F., Olsen, A., Lucas, S., Elkin, C.,
Uberbacher, E., Frazier, M., et al. (2001) Initial sequencing and analysis of the human
genome. Nature 409, 860-921.
176. Robbins, S. L., Cotran, R. S., and Kumar, V., eds (1994) Pathologic Basis of Disease,
W.B. Saunders Company, Philadelphia, Pennsylvania, 19106.
177. Couser, W. G. (1998) Pathogenesis of glomerular damage in glomerulonephritis. Nephrol
Dial Transplant 13, 10-15.
178. Yamada, K., Hori, Y., Hanafusa, N., Okuda, T., Nagano, N., Choi-Miura, N. H., Couser,
W. G., Miyata, T., Kurokawa, K., Fujita, T., and Nangaku, M. (2001) Clusterin is up-
regulated in glomerular mesangial cells in complement- mediated injury. Kidney Int 59,
137-146.
188 179. Nagy, G., Brozik, M., Varga, L., Fust, G., Kirschfink, M., Kiss, E., and Gergely, P.
(2000) Usefulness of detection of complement activation products in evaluating SLE
activity. Lupus 9, 19-25.
180. Chiu, Y. Y., Nisihara, R. M., Wurzner, R., Kirschfink, M., and de Messias-Reason, I. J.
(1998) SC5b-9 is the most sensitive marker in assessing disease activity in Brazilian SLE
patients. J Investig Allergol Clin Immunol 8, 239-244.
181. Choi, N. H., Mazda, T., and Tomita, M. (1989) A serum protein SP40,40 modulates the
formation of membrane attack complex of complement on erythrocytes. Mol Immunol 26,
835-840.
182. Hochgrebe, T. T., Humphreys, D., Wilson, M. R., and Easterbrook-Smith, S. B. (1999) A
reexamination of the role of clusterin as a complement regulator. Exp Cell Res 249, 13-
21.
183. Quigg, R. J., Lim, A., Haas, M., Alexander, J. J., He, C., and Carroll, M. C. (1998)
Immune complex glomerulonephritis in C4- and C3-deficient mice. Kidney Int 53, 320-
330.
184. Falk, R. J., and Jennette, J. C. (1986) Immune complex induced glomerular lesions in C5
sufficient and deficient mice. Kidney Int 30, 678-686.
185. Frymoyer, P. A., Tatum, A. H., and Hajdu, I. (1988) The role of H-2 in apoferritin-
induced murine immune complex glomerulonephritis. Lab Invest 58, 674-681.
186. Iskandar, S. S., Jennette, J. C., Wilkman, A. S., and Becker, R. L. (1982) Interstrain
variations in nephritogenicity of heterologous protein in mice. Lab Invest 46, 344-351.
189 187. McLeish, K. R., Stelzer, G. T., Gohara, A. F., and Wallace, J. H. (1986) Variable
susceptibility to immune complex glomerulonephritis among mice sharing the same
major histocompatibility complex. Immunol Invest 15, 541-547.
188. Welch, T. R., Frenzke, M., and Witte, D. (2000) Evidence of a role for local complement
expression in a murine model of progressive glomerulonephritis. Pediatr Res 48, 200-
205.
189. Iskandar, S. S., Gifford, D. R., and Emancipator, S. N. (1988) Immune complex acute
necrotizing glomerulonephritis with progression to diffuse glomerulosclerosis. A murine
model. Lab Invest 59, 772-779.
190. Ziccardi, R. J. (1983) The first component of human complement (C1): activation and
control. Springer Semin Immunopathol 6, 213-230.
191. Hory, B., Panouse-Perrin, J., Suzuki, Y., Faivre, R., Saint-Hillier, Y., Coulon, G., and
Leibowitch, J. (1981) [Immune complex nephropathy and hereditary deficiency of C1
esterase inhibitor (author's transl)]. Nouv Presse Med 10, 2193-2196.
192. Hory, B., and Haultier, J. J. (1989) Glomerulonephritis and hereditary angioedema: report
of 2 cases. Clin Nephrol 31, 259-263.
193. Massa, M. C., and Connolly, S. M. (1982) An association between C1 esterase inhibitor
deficiency and lupus erythematosus: report of two cases and review of the literature. J
Am Acad Dermatol 7, 255-264.
194. Pan, C. G., Strife, C. F., Ward, M. K., Spitzer, R. E., and McAdams, A. J. (1992) Long-
term follow-up of non-systemic lupus erythematosus glomerulonephritis in patients with
hereditary angioedema: report of four cases. Am J Kidney Dis 19, 526-531.
190 195. Roeth, P. J., and Easterbrook-Smith, S. B. (1996) C1q is a nucleotide binding protein and
is responsible for the ability of clusterin preparations to promote immune complex
formation. Biochim Biophys Acta 1297, 159-166.
196. Gessner, J. E., Heiken, H., Tamm, A., and Schmidt, R. E. (1998) The IgG Fc receptor
family. Ann Hematol 76, 231-248.
197. Ravetch, J. V., and Bolland, S. (2001) IgG Fc receptors. Annu Rev Immunol 19, 275-290
198. Gomez-Guerrero, C., Duque, N., Casado, M. T., Pastor, C., Blanco, J., Mampaso, F.,
Vivanco, F., and Egido, J. (2000) Administration of IgG Fc fragments prevents
glomerular injury in experimental immune complex nephritis. J Immunol 164, 2092-
2101.
199. Coxon, A., Cullere, X., Knight, S., Sethi, S., Wakelin, M. W., Stavrakis, G., Luscinskas,
F. W., and Mayadas, T. N. (2001) Fc gamma RIII mediates neutrophil recruitment to
immune complexes. a mechanism for neutrophil accumulation in immune-mediated
inflammation. Immunity 14, 693-704.
200. Park, S. Y., Ueda, S., Ohno, H., Hamano, Y., Tanaka, M., Shiratori, T., Yamazaki, T.,
Arase, H., Arase, N., Karasawa, A., Sato, S., Ledermann, B., Kondo, Y., Okumura, K.,
Ra, C., and Saito, T. (1998) Resistance of Fc receptor- deficient mice to fatal
glomerulonephritis. J Clin Invest 102, 1229-1238.
201. Clynes, R., Dumitru, C., and Ravetch, J. V. (1998) Uncoupling of immune complex
formation and kidney damage in autoimmune glomerulonephritis. Science 279, 1052-
1054.
191 202. Marino, M., Ruvo, M., De Falco, S., and Fassina, G. (2000) Prevention of systemic lupus
erythematosus in MRL/lpr mice by administration of an immunoglobulin-binding
peptide. Nat Biotechnol 18, 735-739.
203. Berger, S., Chandra, R., Ballo, H., Hildenbrand, R., and Stutte, H. J. (1997) Immune
complexes are potent inhibitors of interleukin-12 secretion by human monocytes. Eur J
Immunol 27, 2994-3000.
204. Berger, S., Ballo, H., and Stutte, H. J. (1996) Immune complex-induced interleukin-6,
interleukin-10 and prostaglandin secretion by human monocytes: a network of pro- and
anti-inflammatory cytokines dependent on the antigen:antibody ratio. Eur J Immunol 26,
1297-1301.
205. Berger, S., Ballo, H., and Stutte, H. J. (1996) Distinct antigen-induced cytokine pattern
upon stimulation with antibody-complexed antigen consistent with a Th1-->Th2-shift.
Res Virol 147, 103-108.
206. Berger, S., Ballo, H., and Stutte, H. J. (1996) [Immune complexes induce the secretion of
Th2 cytokines in human monocytes]. Verh Dtsch Ges Pathol 80, 288-292.
207. Holdsworth, S. R., Kitching, A. R., and Tipping, P. G. (1999) Th1 and Th2 T helper cell
subsets affect patterns of injury and outcomes in glomerulonephritis. Kidney Int 55, 1198-
1216.
208. Levinson, ed (1996) Medical Microbiology & Immunology: Examination and Board
Review, Appleton & Lange.
209. Horwitz, D. A., Gray, J. D., Behrendsen, S. C., Kubin, M., Rengaraju, M., Ohtsuka, K.,
and Trinchieri, G. (1998) Decreased production of interleukin-12 and other Th1-type
192 cytokines in patients with recent-onset systemic lupus erythematosus. Arthritis Rheum 41,
838-844.
210. Funauchi, M., Ikoma, S., Enomoto, H., and Horiuchi, A. (1998) Decreased Th1-like and
increased Th2-like cells in systemic lupus erythematosus. Scand J Rheumatol 27, 219-
224.
211. Richaud-Patin, Y., Alcocer-Varela, J., and Llorente, L. (1995) High levels of TH2
cytokine gene expression in systemic lupus erythematosus. Rev Invest Clin 47, 267-272.
212. Hagiwara, E., Gourley, M. F., Lee, S., and Klinman, D. K. (1996) Disease severity in
patients with systemic lupus erythematosus correlates with an increased ratio of
interleukin-10:interferon-gamma- secreting cells in the peripheral blood. Arthritis Rheum
39, 379-385.
213. Houssiau, F. A., Lefebvre, C., Vanden Berghe, M., Lambert, M., Devogelaer, J. P., and
Renauld, J. C. (1995) Serum interleukin 10 titers in systemic lupus erythematosus reflect
disease activity. Lupus 4, 393-395.
214. Park, Y. B., Lee, S. K., Kim, D. S., Lee, J., Lee, C. H., and Song, C. H. (1998) Elevated
interleukin-10 levels correlated with disease activity in systemic lupus erythematosus.
Clin Exp Rheumatol 16, 283-288.
215. Tsokos, G. C., Boumpas, D. T., Smith, P. L., Djeu, J. Y., Balow, J. E., and Rook, A. H.
(1986) Deficient gamma-interferon production in patients with systemic lupus
erythematosus. Arthritis Rheum 29, 1210-1215.
216. Ofosu-Appiah, W., Sfeir, G., Viti, D., and Burashnikova, E. (1999) Suppression of
systemic lupus erythematosus disease in mice by oral administration of kidney extract. J
Autoimmun 13, 405-414.
193 217. Penny, M. J., Boyd, R. A., and Hall, B. M. (1998) Mycophenolate mofetil prevents the
induction of active Heymann nephritis: association with Th2 cytokine inhibition. J Am
Soc Nephrol 9, 2272-2282.
218. Doth, M., Fricke, M., Nicoletti, F., Garotta, G., Van Velthuysen, M. L., Bruijn, J. A., and
Gleichmann, E. (1997) Genetic differences in immune reactivity to mercuric chloride
(HgCl2): immunosuppression of H-2d mice is mediated by interferon-gamma (IFN-
gamma). Clin Exp Immunol 109, 149-156.
219. Kang, J., Bai, K. M., Wang, B. L., Yao, Z., Pang, X. W., and Chen, W. F. (1994)
Increased production of interleukin 4 in children with simple idiopathic nephrotic
syndrome. Chin Med J (Engl) 107, 347-350.
220. Huang, X. R., Holdsworth, S. R., and Tipping, P. G. (1997) Th2 responses induce
humorally mediated injury in experimental anti- glomerular basement membrane
glomerulonephritis. J Am Soc Nephrol 8, 1101-1108.
221. Wilder, R. L. (1998) Hormones, pregnancy, and autoimmune diseases. Ann N Y Acad Sci
840, 45-50.
222. Ostensen, M. (1999) Sex hormones and pregnancy in rheumatoid arthritis and systemic
lupus erythematosus. Ann N Y Acad Sci 876, 131-143; discussion 144.
223. Carlsten, H., Nilsson, N., Jonsson, R., Backman, K., Holmdahl, R., and Tarkowski, A.
(1992) Estrogen accelerates immune complex glomerulonephritis but ameliorates T cell-
mediated vasculitis and sialadenitis in autoimmune MRL lpr/lpr mice. Cell Immunol 144,
190-202.
194 224. Wu, W. M., Lin, B. F., Su, Y. C., Suen, J. L., and Chiang, B. L. (2000) Tamoxifen
decreases renal inflammation and alleviates disease severity in autoimmune NZB/W F1
mice. Scand J Immunol 52, 393-400.
225. Giron-Gonzalez, J. A., Moral, F. J., Elvira, J., Garcia-Gil, D., Guerrero, F., Gavilan, I.,
and Escobar, L. (2000) Consistent production of a higher TH1:TH2 cytokine ratio by
stimulated T cells in men compared with women. Eur J Endocrinol 143, 31-36.
226. Wang, Y., Hu, Q., Madri, J. A., Rollins, S. A., Chodera, A., and Matis, L. A. (1996)
Amelioration of lupus-like autoimmune disease in NZB/WF1 mice after treatment with a
blocking monoclonal antibody specific for complement component C5. Proc Natl Acad
Sci U S A 93, 8563-8568.
227. Brandt, J., Pippin, J., Schulze, M., Hansch, G. M., Alpers, C. E., Johnson, R. J., Gordon,
K., and Couser, W. G. (1996) Role of the complement membrane attack complex (C5b-9)
in mediating experimental mesangioproliferative glomerulonephritis. Kidney Int 49, 335-
343.
228. Nangaku, M. (1998) Complement regulatory proteins in glomerular diseases. Kidney Int
54, 1419-1428.
229. Abe, K., Miyazaki, M., Koji, T., Furusu, A., Ozono, Y., Harada, T., Sakai, H., Nakane, P.
K., and Kohno, S. (1998) Expression of decay accelerating factor mRNA and
complement C3 mRNA in human diseased kidney. Kidney Int 54, 120-130.
230. Arora, M., Arora, R., Tiwari, S. C., Das, N., and Srivastava, L. M. (2000) Expression of
complement regulatory proteins in diffuse proliferative glomerulonephritis. Lupus 9, 127-
131.
195 231. Nangaku, M., Alpers, C. E., Pippin, J., Shankland, S. J., Kurokawa, K., Adler, S.,
Morgan, B. P., Johnson, R. J., and Couser, W. G. (1998) CD59 protects glomerular
endothelial cells from immune-mediated thrombotic microangiopathy in rats. J Am Soc
Nephrol 9, 590-597.
232. Matsuo, S., Nishikage, H., Yoshida, F., Nomura, A., Piddlesden, S. J., and Morgan, B. P.
(1994) Role of CD59 in experimental glomerulonephritis in rats. Kidney Int 46, 191-200.
233. Martyn, P., Charlesworth, J. A., Peake, P. W., and Pussell, B. A. (1990) Studies of the
acute phase response in experimental serum sickness. J Clin Lab Immunol 31, 69-73.
234. Cunningham, P. N., Holers, V. M., Alexander, J. J., Guthridge, J. M., Carroll, M. C., and
Quigg, R. J. (2000) Complement is activated in kidney by endotoxin but does not cause
the ensuing acute renal failure. Kidney Int 58, 1580-1587.
235. Pepys, M. B., Rogers, S. L., and Evans, D. J. (1982) Role of the acute phase response in
the Shwartzman phenomenon. Clin Exp Immunol 47, 289-295.
236. Schiller, B., Cunningham, P. N., Alexander, J. J., Bao, L., Holers, V. M., and Quigg, R. J.
(2001) Expression of a soluble complement inhibitor protects transgenic mice from
antibody-induced acute renal failure. J Am Soc Nephrol 12, 71-79.
237. Sekine, H., Reilly, C. M., Molano, I. D., Garnier, G., Circolo, A., Ruiz, P., Holers, V. M.,
Boackle, S. A., and Gilkeson, G. S. (2001) Complement component C3 is not required
for full expression of immune complex glomerulonephritis in MRL/lpr mice. J Immunol
166, 6444-6451.
238. Suzuki, Y., Tsukushi, Y., Ito, M., and Nagamatsu, T. (1987) Antinephritic effect of Y-
19018, a thromboxane A synthetase inhibitor, on crescentic-type anti-GBM nephritis in
rats. Jpn J Pharmacol 45, 177-185.
196
239. Izumino, K., Iida, H., Asaka, M., and Sasayama, S. (1987) Effect of the antiplatelet
agents ticlopidine and dipyridamole on nephrotoxic serum nephritis in rats. Nephron 45,
306-310.
240. Jacquier-Sarlin, M. R., Fuller, K., Dinh-Xuan, A. T., Richard, M. J., and Polla, B. S.
(1994) Protective effects of hsp70 in inflammation. Experientia 50, 1031-1038.
241. Polla, B. S., and Cossarizza, A. (1996) Stress proteins in inflammation. Exs 77, 375-391
242. Joyeux, M., Godin-Ribuot, D., Yellon, D. M., Demenge, P., and Ribuot, C. (1999) Heat
stress response and myocardial protection. Fundam Clin Pharmacol 13, 1-10
243. Hightower, L. E., Brown, Renfro, J. L., Perdrizet, G. A., Rewinski, M., Guidon, P. T., Jr.,
Mistry, T., and House, S. D. (2000) Tissue-level cytoprotection. Cell Stress Chaperones
5, 412-414.
244. Ianaro, A., Ialenti, A., Maffia, P., Pisano, B., and Di Rosa, M. (2001) HSF1/hsp72
pathway as an endogenous anti-inflammatory system. FEBS Lett 499, 239-244.
245. Schett, G., Redlich, K., Xu, Q., Bizan, P., Groger, M., Tohidast-Akrad, M., Kiener, H.,
Smolen, J., and Steiner, G. (1998) Enhanced expression of heat shock protein 70 (hsp70)
and heat shock factor 1 (HSF1) activation in rheumatoid arthritis synovial tissue.
Differential regulation of hsp70 expression and hsf1 activation in synovial fibroblasts by
proinflammatory cytokines, shear stress, and antiinflammatory drugs. J Clin Invest 102,
302-311.
246. Santoro, M. G. (2000) Heat shock factors and the control of the stress response. Biochem
Pharmacol 59, 55-63.
197 247. Donnelly, T. J., Sievers, R. E., Vissern, F. L., Welch, W. J., and Wolfe, C. L. (1992) Heat
shock protein induction in rat hearts. A role for improved myocardial salvage after
ischemia and reperfusion? Circulation 85, 769-778.
248. Hotchkiss, R., Nunnally, I., Lindquist, S., Taulien, J., Perdrizet, G., and Karl, I. (1993)
Hyperthermia protects mice against the lethal effects of endotoxin. Am J Physiol 265,
R1447-1457.
249. Koh, Y., Lim, C. M., Kim, M. J., Shim, T. S., Lee, S. D., Kim, W. S., Kim, D. S., and
Kim, W. D. (1999) Heat shock response decreases endotoxin-induced acute lung injury in
rats. Respirology 4, 325-330.
250. Perdrizet, G. A. (1995) Heat shock and tissue protection. New Horiz 3, 312-320.
251. Leppa, S., and Sistonen, L. (1997) Heat shock response--pathophysiological implications.
Ann Med 29, 73-78.
252. Hutter, J. J., Mestril, R., Tam, E. K., Sievers, R. E., Dillmann, W. H., and Wolfe, C. L.
(1996) Overexpression of heat shock protein 72 in transgenic mice decreases infarct size
in vivo. Circulation 94, 1408-1411.
253. Coronato, S., Di Girolamo, W., Salas, M., Spinelli, O., and Laguens, G. (1999) [Biology
of heat shock proteins]. Medicina (B Aires) 59, 477-486.
254. Storm, F. K., Mahvi, D. M., and Gilchrist, K. W. (1996) Heat shock protein 27
overexpression in breast cancer lymph node metastasis. Ann Surg Oncol 3, 570-573.
255. Nanbu, K., Konishi, I., Mandai, M., Kuroda, H., Hamid, A. A., Komatsu, T., and Mori, T.
(1998) Prognostic significance of heat shock proteins HSP70 and HSP90 in endometrial
carcinomas. Cancer Detect Prev 22, 549-555.
198 256. Uozaki, H., Ishida, T., Kakiuchi, C., Horiuchi, H., Gotoh, T., Iijima, T., Imamura, T., and
Machinami, R. (2000) Expression of heat shock proteins in osteosarcoma and its
relationship to prognosis. Pathol Res Pract 196, 665-673.
257. Lambot, M. A., Peny, M. O., Fayt, I., Haot, J., and Noel, J. C. (2000) Overexpression of
27-kDa heat shock protein relates to poor histological differentiation in human
oesophageal squamous cell carcinoma. Histopathology 36, 326-330.
258. Hoang, A. T., Huang, J., Rudra-Ganguly, N., Zheng, J., Powell, W. C., Rabindran, S. K.,
Wu, C., and Roy-Burman, P. (2000) A novel association between the human heat shock
transcription factor 1 (HSF1) and prostate adenocarcinoma. Am J Pathol 156, 857-864.
259. Cahill, C. M., Waterman, W. R., Xie, Y., Auron, P. E., and Calderwood, S. K. (1996)
Transcriptional repression of the prointerleukin 1beta gene by heat shock factor 1. J Biol
Chem 271, 24874-24879.
260. Singh, I. S., Viscardi, R. M., Kalvakolanu, I., Calderwood, S., and Hasday, J. D. (2000)
Inhibition of tumor necrosis factor-alpha transcription in macrophages exposed to febrile
range temperature. A possible role for heat shock factor-1 as a negative transcriptional
regulator. J Biol Chem 275, 9841-9848.
261. Lee, J. E., Yenari, M. A., Sun, G. H., Xu, L., Emond, M. R., Cheng, D., Steinberg, G. K.,
and Giffard, R. G. (2001) Differential neuroprotection from human heat shock protein 70
overexpression in in vitro and in vivo models of ischemia and ischemia- like conditions.
Exp Neurol 170, 129-139.
262. Packman, C. H., Rosenfeld, S. I., and Leddy, J. P. (1985) High-density lipoprotein and its
apolipoproteins inhibit cytolytic activity of complement. Studies on the nature of
inhibitory moiety. Biochim Biophys Acta 812, 107-115.
199