Molecular Biology in Prostate Cancer

Contents:

These stories are about genes and proteins in cells, and how they malfunction to make cancer.
This web page links to summaries below, and then to the original stories.
Metastatic transformation. Thymosin beta-15 makes cancer cells motile. Caveolin contributes another step to metastases. Endothelin stimulates cancer growth, and a drug blocks its receptor.
Hormone resistance. Cadherin is a brake on cell growth that fails in metastasis. PTEN is another brake. HER-2/neu and PI3 are accelerators. AKT is an oncogene. p53 is the master brake, but bcl-2 overrides it. Ribozymes inactivate bcl-2.
Growth factors. Cancer cells need IGF to survive, but an enzyme inactivates IGF in patients.
Angiogenesis. Angiogenesis inhibitors such as angiostatin and endostatin "regress every tumor tested so far in animals." What about humans?
Monoclonal antibody therapy. MABs to PSMA and HER-2/neu get immune responses in Phase II trials.
Tumor markers. The breast cancer gene BRCA2 isn't associated with PCa, but another gene nearby on chromosome 13 is. So are genes on chromosome 6 So is DD3, on chromosome 9. PSCA is on chromosome 8, a hot spot for cancer mutations.

Here is an Introductory background to the molecular biology in these stories.

Metastatic transformation

Thymosin beta-15 is an oncogene that makes cancer cells motile. Caveolin is another oncogene that contributes another step to metastases. Endothelin stimulates cancer growth generally, and a receptor blocking compound was being tested to treat PCa.

Thymosin beta-15
New protein causes motility;
Assay differentiates middle Gleason grades

Thymosin beta-15 is an oncogene that upregulates cell motility in prostate carcinoma cells, said Bruce R. Zetter, PhD, Harvard, in Nature Medicine, Dec. 1996. Cell lines with thymosin beta-15 were mobile and metastatic, while non-metastatic lines were not. Highly invasive pathology specimens were positive, and non-invasive cells were usually not. Single cells invading stroma were intensely stained by an antibody stain. Staining correlated with Gleason grade, and may predict clinical outcome better than Gleason grade. Thymosin beta-15 binds to actin. KAI-1, in contrast, suppresses motility.

Caveolin
Second PCa metastasis gene found
Structural protein locates signal transduction

Caveolin is another oncogene that is expressed in metastatic prostate cancer cells, said Timothy C. Thompson, PhD, Baylor, at the 4th CapCURE scientific retreat in Lake Tahoe, 1997. It was found in metastatic tumors in p53-knockout mice that were transformed by a retrovirus. It was also found in 43/51 prostatectomy specimens, and in breast cancer. Caveolin protein is a structural protein of cavaolae, organelles in the trans-Golgi network, and is involved in the localization of signal transduction molecules.

Endothelin
ET-1 receptor methylated in metastases;
can a drug block its effects?

Cancer is an imbalance between the cell's accelerator and brake. Endothelin-1 is the accelerator. A methyl group is the switch that turns it on. ET-1 is elevated in men with metastatic prostate cancer, said Joel B. Nelson, MD, Johns Hopkins, at the AUA's 92nd annual meeting in New Orleans, 1997. He found ET-1 in every human prostate cancer cell line tested, 14/14 primary cancers, and 14/16 metastatic sites. ET-1 induces PCa proliferation directly, and enhances IGF-I IGF-II, PDGF, BGF, and EGF in vitro. ET-1 stimulates osteoblasts and inhibits osteoclasts. ET-1 has 2 receptors on the prostate--ETA and ETB. In PCa, ETA is unchanged. But in 70% of PCa, the ETB gene is methylated, which blocks its expression, and ETB receptors are reduced to undetectable levels, which may increase ET-1. An ETA blocking compound, Abbot ABT 627, inhibited ET-1 in vitro, and is in Phase I trials, currently recruiting patients, by Michael A. Carducci, Johns Hopkins. (Another gene, GST-pi, is also methylated in almost every prostate cancer.) The 30% of PCa in which ETB is not methylated could have the same mutation by a different mechanism, since the gene is located on chromosome 13q, "where allelic loss is extremely common," said Nelson, near BRCA2, RB1, and DBM.

Hormone resistance

Cancer is an imbalance between the cell's accelerator and its brakes. Cadherin is one of the brakes, and an abnormal form is found in metastatic prostate cells. PTEN is another brake, and is missing in many cancers. HER-2/neu and PI3 are accelerators. AKT is another oncogene. p53 is one of the important brakes, but bcl-2 overrides it. Ribozymes are RNA enzymes which may inactivate bcl-2.

Cadherin
Membrane binding protein needed for apoptosis,
altered cadherin lets it fall into nucleus

T6, a mutation in the cadherin gene family, confers hormone resistance to LNCaP cells, and is elevated in metastatic primary prostate cancer specimens, said Ralph Buttyan, MD, Columbia U., at the 93rd AUA meeting in San Diego, 1998. Cadherins are located on the cell membrane, and bind cells to each other. Cadherins release the cells from each other under 2 circumstances: apoptosis or mitosis. "In order for the cells to divide, they first have to separate from their neighboring cells," said Buttyan. Cadherins are complexed with another protein, beta-catinin, which is needed at the cell membrane for apoptosis. But the mutated T6 cadherin doesn't hold catinins as well as the normal cadherin, and it lets beta-catinin fall into the nucleus, where it no longer causes the cell to undergo apoptosis.

Separately, Eric D. Schwab, PhD, in the laboratory of Kenneth J. Pienta, MD, U. Michigan, found 2 new DNA sequences, rPMET-1 and rPMET-2, when the AT.1 Dunning rat prostate cancer line becomes metastatic. They found a similar gene with a 700-base-pair match in LNCaP and other lines, and are now trying to show that rPMET-1 and rPMET-2 indeed cause metastases.

PTEN, Her-2/neu
Androgen-dependent cells not dead, just resting,
unless they have a HER-2/neu mutation

Cancer is an imbalance between the cell's accelerator and brake. HER-2/neu, a breast cancer gene, stimulates the cell even when its natural accelerator, androgen, is turned off. PTEN is the brake.

Her-2/neu can make prostate cancer cells androgen-independent, said Charles L. Sawyers, MD, UCLA, at the 5th CapCURE retreat, 1998. Cells become androgen-resistant when alternate signaling pathways bypass the androgen pathway. HER-2/neu "sneaks around" the androgen receptor, and "talks with" downstream molecules, stimulating growth even without androgen. The HER-2/neu gene, at 17q21, is overproduced in some prostate cancers. "When we put HER-2/neu into LNCaP and LAPC-4 cells it made them androgen-independent," said Sawyers. "It did so by turning on the androgen receptor."

HER-2/neu is a tyrosine kinase, which adds a phosphate group to the tyrosine amino acid of certain proteins, to turn them on.

PTEN, at 10q23, is a phosphatase, which removes the phosphatase group, to turn them off. 10q23 is deleted or unexpressed in 60% of advanced prostate cancers. PTEN is "in my opinion the most commonly mutated gene in advanced prostate cancer," said Sawyers.

Separately, Sawyers found evidence that LAPC-9 cells were dependent on androgen for growth, but not survival. If he injected a small number of LAPC-9 cells sc into castrated SCID mice, the cells remained dormant for 110 days, they survived, and then after addition of androgen, grew into a tumor.

PTEN
Newly-discovered tumor suppressor,
controls PI3 kinase/AKT and other pathways

William R. Sellers, MD, Dana-Farber, described the PI3 kinase/AKT pathway. at the 5th CapCURE scientific retreat, 1998. AKT is the accelerator, PTEN is the brake. They turn cancer cells on and off with a phosphate switch. AKT is an oncogene. AKT is regulated up by PI3 kinase, which adds a phosphate group to a lipid intermediary to turn it on. AKT is regulated down by PTEN, which removes the phosphate group from the lipid to turn it off. The "important finding," said Sellers, was that "some PTEN mutations would remove phosphate groups from proteins but not from lipids."

The tumor suppressor gene PTEN, at chromosome 10q23, must be important because it's mutated in a wide range of tumors, said Sellers. PTEN is inactivated in LNCaP, where it had a 2-base-pair deletion, and DU145, where the protein had the amino acid Met substituted for Leu. When cells turn cancerous, PTEN sometimes kills them by apoptosis, but sometimes just stops cell growth and leaves them quiescent. PTEN is found in 15% of primary tumors, "but it may have significance for a much larger number," said Sellers, "because it might lead us to abnormalities in other places on that pathway."

bcl-2
Apoptosis is critical path to cancer;
therapeutic targets emerge

Cancer is an imbalance between the cell's accelerator and brakes, said John C. Reed , MD, PhD, Burnham Institute, at the 5th CapCURE retreat, 1998. p53 is the brakes. bcl-2 overrides the brakes. Prostate epithelial cells should die when their androgens are cut off, but they continue growing because of defects in the steroid hormone pathway, mitochondria pathway, the upstream path, the downstream path, and kinases that prevent apoptosis. About half of hormone-refractory prostate cancers express bcl-2. When bcl-2 is overexpressed, cells resist apoptosis from hormone deprivation, radiation, hypoxia, the immune system, and p53 generally. "We're particularly interested in apoptosis, because the only good tumor cell is a dead tumor cell," said Reed. "The genetic instability in tumor means that if you don't kill them, they're going to find a way around almost any therapy." The immune system merely activates the apoptotic pathway, so if the apoptotic pathway is defective, the immune system can't kill it anyway.

bcl-2
Gene makes PCa hormone-refractory
but RNA catalyst destroys it

bcl-2 protects cells from apoptosis, said Ralph Buttyan, MD, Columbia U., at the 92nd AUA annual meeting, New Orleans, 1997. When bcl-2 is absent, cells die; when it is overexpressed, cells ignore apoptosis signals. Buttyan's group has developed a ribozyme (1, 2) an RNA molecule which destroys bcl-2 mRNA in vivo and in vitro. "The beauty of a ribozyme," said Eric Goluboff, MD, Columbia U., "is that a single molecule of ribozyme can cleave hundreds or thousands of mRNAs, whereas an antisense DNA can only bind to one mRNA." When the ribozyme is transfected into normal LNCaP cells, with low bcl-2, it kills them. When it is transfected into refractory prostate cancer cells with high bcl-2, it makes them sensitive again to radiation or chemotherapy. They hope to transmit the ribozyme into mice with an adenovirus vector, "and effectively cure the hormone resistant prostate cancer in these mice," said Buttyan.

Growth factors

Cancer cells need IGF to survive, but an enzyme can inactivate IGF, and some patients have done well in Phase I trials.

IGF-1
Receptor blocker cures 2 glioblastoma patients,
anti-sense sequence kills PCa cells in culture too

Normal cells don't need insulin-like growth factor to survive, but cancer cells do. When the IGF pathway is inhibited, normal cells stop growing and become dormant, but "tumor cells die immediately," said Renato Baserga, MD, Thomas Jefferson U, at the 5th CapCURE retreat, 1998. IGF-1 is required for anchorage-independent growth. In a Phase 1 study of terminal malignant glioblastoma, Baserga destroyed the IGF-1 receptor. He treated 10 patients with an oligonucleotide that targets IGF-1 mRNA. 2 patients had a "dramatic" long-term response.

Another way to interrupt the IGF-1 pathway is with decoy receptors that compete for IGF-1 with the tumor cell receptors. One receptor, 486/STOP, caused growth inhibition and apoptosis of 5 human tumor cell lines. More important, it killed neighboring cells with a bystander effect. There is some evidence that, after most of the tumor cells are killed by IGF-1, the immune system destroys the rest.

Angiogenesis

Angiogenesis inhibitors such as angiostatin and endostatin "can regress every tumor tested so far in animals," including PCa, but it's hard to synthetize enough to treat humans.

Endostatin
Angiogenisis inhibitors stop tumors permanently;
many twists in scaleup from mice to man

"The hallmark of these specific angiogenesis inhibitors continues to be lack of acquired drug resistance, and lack of detectable toxicity," said Judah Folkman, MD, Harvard, at the 5th CapCURE scientific retreat in 1998. Even after treatment was stopped, they produced a "self-sustained dormancy that is indefinite." The latest generation of angiogenesis inhibitors, including angiostatin and endostatin, "can regress every tumor tested so far in animals." LNCaP cells were planted in the mouse prostate. In untreated mice, the PSA went over 100, the tumors fill the pelvis, and push up the liver. Other mice were treated at about PSA 40, when the tumor was just coming out of the pelvis, and the tumor was "simply eradicated." One set of mice has been followed for 293 days. They could make enough endostatin to treat mice, but when they tried to scale it up for enough to treat humans, it didn't work. They found out that the medium was zinc-deficient in scaleup.

Folkman noted that as they get old enough, most women have microscopic dormant breast cancer, most men develop microscopic dormant prostate cancer, and most men and women develop microscopic thyroid cancer, the majority of which are never clinically significant. They are avascular tumors, said Folkman. He suggested that they grow to the avascular limit and stay there until angiogenesis is switched on.

Monoclonal antibody therapy

PSMA antibody
Monoclonal antibodies work for other cancers,
seem ideal for prostate cancer too

Prostate cancer is "ideally suited" for monoclonal antibody therapy, said Neil Harrison Bander, MD, New York Hospital-Cornell. "Prostate cancer metastasizes predominantly to bone marrow and lymph nodes," sites that get high levels of circulating antibodies, and have been effectively treated by antibodies in other settings. PCa has a small volume of tumor. The antibody target has been prostate-specific membrane antigen, first with Prost 30, the Prostascint antibody, and now with J591.

In a Phase I trial, 19 patients got Prost 30. "Serendipidously and unexpectedly" 4/7 (57%) evaluable patients had PSA declines of greater than 50%. In a Phase I/II trial in 21 patients with rising PSAs, 23% had PSA declines of greater than 50%. 10 patients at high risk following radical prostatectomy got Prost 30, and after 2.4 years, had PSA failure of 30%, compared to expected 60%. The PSMA variant found in PCa spans the cell membrane, with most of it external. Pros 30 targets the internal part, and can bind only to dead cells, but J591 targets the external part, and can bind to living cells. PSMA is also expressed by tumor vascular endothelial cells, which raises the possibility of a "double whammy," said Bander. The mouse version of the antibody is starting Phase I trials.

HER2/neu antibody
Bispecific antibodies shrunk a few tumors,
randomized trial for PCa starting

5-year Phase II trials are starting for a monoclonal antibody designed to stimulate an immune response to prostate cancers expressing HER-2/neu oncogene, said Aaron E. Katz, MD, Columbia U., at the 4th CapCURE scientific retreat, 1997. Radical prostatectomy patients with tumor at the margin or seminal vesicle involvement will get 6 weekly infusions of humanized mouse antibody. The bispecific antibody responds to 2 idiotypes: (1) HER-2/neu protein, a tyrosine kinase receptor on the cell surface (2) Fc receptor on monocytes, macrophages, and activated neutrophils, which triggers killing. The antibody brings 2 cells together, and binds to an adhesion molecule on killer T cells that activates T cells to destroy the cancer cell. The antibody, MDX-210, was developed by David Segal, MD, NCI, and manufactured by Medarex, Inc., Annendale, NJ. In Phase II studies in kidney cancer, MDX-210 reduced metastases in 2/4 kidney cancer patients, and lowered PSA in 1 prostate cancer patient, according to Philip Atherton, MD, U. Birmingham, UK.

Tumor markers

The breast cancer gene BRCA2 is not associated with PCa, but another gene nearby on chromosome 13 is. Missing genes on chromosome 6 cause worse pathological stage and recurrence. A new gene, DD3, was found on chromosome 9. Another gene, PSCA, was found on chromosome 8, which is a hot spot for cancer mutations.

BRCA2, 6q LOH, DD3
Disrupted spots on 3 chromosomes
hint at prostate cancer genes

Men who inherit the breast cancer gene, BRCA2 are 3 times as likely to have prostate cancer, but BRCA2 is not responsible, said Chunde Li, MD, PhD, Karolinska Hospital, Stockholm, at the 92nd AUA meeting in New Orleans, 1997. BRCA2 is in a genetic hot spot, at the same region of the chromosome, 13q12-14, as the retinoblastoma (RB1) and the disrupted in B cell malignancy (DBM) gene. Li examined 35 prostate cancers, 25 from primary tumors, 6 from lymph nodes and 4 from brain metastases. 17/35 (49%) displayed loss of heterozygosity. But BRCA2 was not correlated with prostate cancer. RB1 and DBM were frequently deleted, as were 8 standard PCR markers, but were not correlated with grade, stage or survival. So there must be an unknown deletion that causes prostate cancer in that region, said Li.

Vasantha Srikantan, PhD, and Judd W. Moul, of the Uniformed Services University of Health Sciences, analyzed 10 markers on chromosome 6q in samples from 37 patients to find which markers were missing. 10 patients had LOH, and the losses were correlated with pathologic stage and recurrence.

Marion J.G. Bussemakers, U. Hospital, Nijmegen, Netherlands, found a new gene, DD3, specific to prostate cancer, on 9q21-22.

PSCA
New prostate cancer gene found--
Does cancer start in basal cell epithelium?

Robert E. Reiter, MD, UCLA, found a new gene, prostate stem cell antigen (PSCA), at 8q24.2, with a 123-amino acid protein of unknown function expressed on the surface of prostate basal epithelial cells, he reported in PNAS, 17 February 1998. "It may be more specific for cancer than PSA," said Reiter. But it didn't correlate with stage, grade or outcome. A mouse protein had a 70% similar sequence, and seems to be a mouse version of PSCA. A human protein, stem cell antigen 2, with 30% similarity, seems to prevent apoptosis of developing T cells in the thymus. 8q is a hot spot that often gains sequences in cancer; 8p often loses sequences.