The Accidental Therapist
|Courtesy of USC|
The idea for his vector came during a conversation with Gordon about the need for a gene therapy vector that would accumulate in tumors after being injected intravenously. Anderson had been calling for such a vector for years. Recalling work he'd done on von Willebrand's factor, Hall tossed out a top-of-the-head remark intended as a joke. "I could design a vector so it would stick to collagen," he remembers laughingly saying, "but who wants to transduce straw?" His jest unexpectedly earned not laughter but silence. In a pause of reflection, it dawned on both of them that targeting collagen might be uncommonly useful.
In healthy tissue collagen is buried under cells, proteins and layers and is exposed only in instances of tissue disruption. "Every lesion of every major disease has disruption of collagen," Hall notes, adding that for cancer the connection to collagen is threefold. Cancer brings about new blood vessel formation (angiogenesis), exposing collagen through turnover of the extracellular matrix. Second, collagen is exposed as tumors edge their way into tissues through the microscopic digestive actions of secreted metalloproteases. Third, one of the body's reactions to a tumor is an attempt to wall it off with a connective tissue capsule; this too exposes collagen.
The body's surveillance system for wounds depends on recognition of exposed collagen by coagulation factor VIII, also known as von Willebrand's factor. "If you cut yourself, platelets plug the hole," explains Hall, and the platelets are guided by von Willebrand's factor, which circulates in the bloodstream until it finds and binds collagen exposed by wounds. One of its domains binds collagen and another domain binds a platelet. Mutations in the protein impair clot formation; von Willebrand's disease, caused by a dominant mutation of the protein, is a form of hemophilia.
Hall's innovation was to add the collagen-binding domain of von Willebrand's factor to a retroviral vector, creating a vector with a surveillance function for wounds, including lesions caused by tumors.1 He accomplished this by grafting the domain onto a retroviral vector's envelope protein. Because the envelope protein is responsible for vector entry into a cell, successful genetic engineering demanded a gain-of-function operation, adding the new without harming the old. Through trial and error Hall added an effective collagen-binding domain to the envelope protein of a Moloney Murine Leukemia Virus (MLV)-based vector without compromising efficiency of cell entry.
The binding power of the domain Hall borrowed is tremendous, considering that despite a high degree of momentum, platelets halt wherever they find exposed collagen. Putting that domain on a retrovirus may seem a bit of overkill, considering how different platelets and retroviruses are in mass--like using an 18-wheeler's brakes to stop a skateboard. But it works. Collagen binding causes the vector to accumulate in a tumor, assuring efficient infection of cancer cells in proximity. Using a retroviral vector adds a safety factor: The therapeutic gene it delivers integrates only in dividing cells--cancer cells, mostly. Quiescent cells within the lesion are unharmed.
A New Death Gene
"She'd [Gordon] had too many patients die in her arms to let a new form of therapy go undeveloped," says Hall. Gordon encouraged Hall's efforts to make therapeutic constructs of the gene. Eventually Hall found that a dominant negative cyclin G1 mutant killed cells more surely than an antisense gene construct, although its mechanism of action still isn't entirely clear. The mutant also had an attractive safety factor: It induced apoptosis only if expressed during the G1 phase of the cell cycle. It had no effect in quiescent cells.
His vector work was still off in the future, but already Hall's commitment to therapy was strong. "There used to be a day when a biochemist could discover a gene or enzyme pathway, write about it, talk about it maybe, put it in the library, and go home," he says. "Probably under the encouragement of Maria even more than French, I was given the obligation that discovery needs to be translated for the benefit of patients." And since gene therapy provides the tools, "all of a sudden you can't go home."
The gene delivery vehicle for the clinical trial will be a retroviral particle, distinct from a retrovirus in that its core packages no viral genes, and so cannot be infectious. The only delivered gene will be the cancer cell killer, mutant cyclin G1. Hall states that the immune response to the particle is weak enough that it can be injected into a patient multiple times.
In one of their preclinical tests, the vector was infused into the livers of nude mice with liver metastases caused by liver-injected human cancer cells. They found the tumors shrank more than 10-fold in surface area, compared to controls.2 Histological analysis showed the metastases had been replaced by little clusters of Kupffer's cells, macrophage-like cells that moved in to clean up the debris left by cancer cell death. The surrounding cells were perfectly pink and healthy.
A second test implanted human tumor cells subcutaneously in the legs of nude mice, and then injected the vector into tail veins. Again, it worked very well.3 Tumor volumes increased an average of 10-fold in controls but stopped growing in vector-treated mice. The control mice died. All vector-treated mice lived.
"Realize where this vector has gone," Hall enthusiastically explains, tracing its circuit to emphasize its power: "Through the heart, through the lungs, all through the capillary beds, then back to the heart." Only a small portion reaches the leg in any one pass. Without the collagen-binding domains, "you would be hard pressed to see any of it in the tumor--but we saw that the vector accumulated within the tumor within an hour." After a series of injections, Hall says that 40 percent of the tumor cells carry the new gene, what he calls an extraordinary mark. "You would have to have a tumor on the surface, like a melanoma, and inject it directly a dozen times to get what we're able to do intravenously."
Hall thinks even better results are possible by delivering different genes. He and Gordon are currently testing vectors that deliver cytokine genes to tumor lesions. Tumor-secreted cytokines like IL-2 and GM-CSF will recruit antitumor lymphocytes and antigen-presenting cells into the lesions. This could be a safer way to use cytokines than simply injecting them into patients, because injections require high doses that are accompanied by considerable toxicity.
One of the most exciting uses Hall foresees for lesion-targeted gene therapy is for preventive surveillance, killing micrometastases that no doctor can see: "Surgeons come to me and say, 'Do you know that we see angiogenic disturbances in the lungs a year before breast cancers will be measurable?'" Sophisticated tracers can reveal blood vessels growing to feed metastatic little nests of breast tumor cells too tiny to be visible. "Those are the kind of things that we can chase," comments Hall.
Because this vector homes in on exposed collagen within all kinds of lesions, with different therapeutic genes there could be other areas of application, including ischemia, stroke, inflammation, and arthritis. But those must wait as the priority goes to diseases where prognoses are poor and treatment alternatives are exhausted.
The phase I clinical trial will test Hall's vector and death gene on liver metastases of colon cancer patients. Adopting a conservative approach, the vector will be delivered directly into the liver via hepatic artery injection; trying delivery from farther away will be left for a later trial.
Funding for the phase I study is available, but where the dollars will come from after that, assuming the first test succeeds, is uncertain. This lack of continued funding demonstrates how unsettled gene therapy still is, a year and a half after the uproar over the accidental death of a patient in a gene therapy clinical trial, the first death in the field's history, even following recent upbeat reports of gene therapy that successfully treated some rare diseases. Novartis Pharma, which has funded much of Hall's work, shies away from a commitment to finance later clinical trials. Negotiations are ongoing between Novartis and USC to set up a new gene therapy company, not yet officially named, responsible for funding phase II and III trials.
In preparation for the clinical trial, Hall and colleagues are busy solving the problems of manufacturing the retroviral particle supernatants that will go into patients. It isn't as if this is without intellectual challenge, but Hall sounds wistful about the less applied pursuits he has left behind. Each task he undertakes to bring lesion-targeted gene therapy to the bedside puts more distance between him and the ivory tower. Biochemist, gene therapist, manufacturer--He's afraid to think about where all this is heading: "Next week I'll be janitor."
1. F.L. Hall et al., "Molecular engineering of matrix-targeted retroviral vectors incorporating a surveillance function inherent in von Willebrand factor," Human Gene Therapy, 11: 983-93, 2000.
2. E.M. Gordon et al., "Inhibition of metastatic tumor growth in nude mice by portal vein infusions of matrix-targeted retroviral vectors bearing a cytocidal cyclin G1 construct," Cancer Research, 60:3343-7, 2000.
3. E.M. Gordon et al. "Systemic administration of a matrix-targeted retroviral vector is efficacious for cancer gene therapy in mice," Human Gene Therapy, 12:193-204, 2001.