The Bricks and Mortar of Personalized Medicine

With an increasing number of therapies derived from a patient's own cells, companies now have experience to draw on in designing commercial infrastructures for delivering these un-drug-like products. Indeed, armed with some small experience around how well they fit--or don't fit--into standard medical practice, cell-based therapies may finally be going commercial. And not only in terms of a transition from acute to chronic wound healing indications, where commercial success was first anticipated, but also as stem cell therapies, which would open up a multi-billon dollar opportunity in regenerative medicine, including replacements for organ and transplant surgery and cures for diabetes and neurological diseases such as Parkinson's Disease and Alzheimer's Disease, as well as autologous cancer immunotherapies. These companies are finding value in the experiences of earlier, technology-constrained companies whose attempts at developing cell therapies failed--Applied Immune Sciences and Systemix Inc., for example, both of which have virtually disappeared. Companies today are trying to overcome the challenges of product preparation, process standardization, and delivery. Other first generation cell therapy companies, like Genzyme Tissue Repair and Cell Genesys Inc., are still around. That a number of senior executives at today's leading-edge cell-therapy companies cut their eye teeth at those first-generation start-ups attests to their continuing belief in the broad commercial opportunities these technologies offer. But significant questions remain: Can companies solve the production economics of delivering products that require cell culture and expansion using a centralized, company-controlled laboratory model? The alternative, a decentralized technology transfer approach, would bring the product closer to the patient, but raises the question of whether the technology exists to standardize processes and satisfy regulators' quality concerns. Then, if these issues are solved, one must ask whether a hospital lab is the right setting for a commercial operation. Are there better commercial infrastructures, such as blood processing centers, with which companies could form alliances? And does the prospect of developing off-the-shelf, allogeneic-cell therapies-mass-produced stem cells that can pass the immune system undetected or vaccines derived from shared tumor antigens from cell lines, for example-threaten to obsolete autologous products before they become profitable?

With an increasing number of therapies derived from a patient's own cells, companies now have experience to draw on in designing infrastructures for delivering these un-drug-like products.

by Mark L. Ratner

• Companies developing autologous therapies have in their sights large markets ranging from connective tissue repair to active cancer immunotherapy to organ replacement and therapies for diabetes and neurological diseases using stem cells.

• Because each production "lot" is patient specific, processes for manufacturing autologous therapies are not scaleable and raise significant new issues of quality control and sample tracking, making production economics an important consideration for developers.

• Culturing cells is also time consuming, often extending the time between a physician's decision to treat and the implementation of therapy to weeks, or even months.

• Given these limitations, must the delivery of this flavor of personalized medicine be made seamless with current medical procedures to be an acceptable business? Or is the potential for improved therapy sufficient to drive change in practice?

• And in many settings, could the prospect of off-the-shelf, allogeneic-cell therapies obsolete autologous products before they become profitable?

Personalized medicine is now widely talked about in terms of genetic profiling, pharmacogenomics, and related plays in diagnostics and drug development. But within the same realm lies a maturing, if not yet established, set of therapies—autologous cell therapies. The common link is that both depend on patient-specific biological material. In the case of pharmacogenomics, that material is information about an individual's genetic make-up. Autologous treatments directly use a patient's own cellular material as their basis.

A small number of these therapies have been around for a while. For example, Epicel, an autologous skin graft product for the treatment of severe burns from Genzyme Tissue Repair (GTR), a division of Genzyme Corp. (and a component of its proposed merger with Biomatrix Inc. that would create Genzyme Biosurgery [See Deal]), has been on the market since the late 1980's. In the mid-90's GTR also introduced Carticel (autologous chondrocytes for treating knee cartilage damage). These initial products, where cells are taken from a patient, expanded in culture, then reintroduced into that patient, may not be stellar successes, but they hint at the potential for autologous cellular therapies.

Indeed, armed with some small experience around how well they fit—or don't fit—into standard medical practice, cell-based medicine may finally be going commercial. And not only in terms of a transition from acute to chronic wound healing indications, where commercial success was first anticipated (see "Wound Healing Regenerates, Slowly,"START-UP, April 1997 [A#1997900063), but also as stem cell therapies, which would open up a multi-billon dollar opportunity in regenerative medicine, including replacements for organ and transplant surgery and cures for diabetes and neurological diseases such as Parkinson's Disease and Alzheimer's Disease (see "Rewriting Fate: The Stem Cell Transplant Business," START-UP, October 2000 [A#2000900165). Autologous cancer immunotherapies—adjuvant-like therapeutic vaccines in which antigens that appear on the surface of a patient's tumor are extracted from a patient and reintroduced in novel forms intended to make them more recognizable to the immune system, in turn stimulating a more active immune response against the tumor—are another area of increasing interest to companies.

And these companies are finding value in the experiences of earlier, technology-constrained companies whose attempts at developing cell therapies failed—applying the lessons those companies learned in trying to overcome the challenges of product preparation, process standardization, and delivery. Companies today believe they are overcoming the technology limitations that hindered their predecessors, which include Applied Immune Sciences (acquired by Rhone-Poulenc Rorer, now a part of Aventis SA [See Deal]) and Systemix Inc., a division of Novartis AG . Both of these companies have virtually disappeared, but others, such as GTR and Cell Genesys Inc. , are still around. That a number of senior executives at today's leading-edge cell-therapy companies cut their eye teeth at those first-generation start-ups attests to their continuing belief in the broad commercial opportunities these technologies offer.

But significant questions remain: Can companies solve the production economics of delivering products that require cell culture and expansion using a centralized, company-controlled laboratory model? The alternative, a decentralized technology transfer approach, would bring the product closer to the patient, but raises the question of whether the technology exists to standardize processes and satisfy regulators' quality concerns. Then, if these issues are solved, one must ask whether a hospital lab is the right setting for a commercial operation. Are there better commercial infrastructures, such as blood processing centers, with which companies could form alliances? And does the prospect of developing off-the-shelf, allogeneic-cell therapies—mass-produced stem cells that can pass the immune system undetected or vaccines derived from shared tumor antigens from cell lines, for example—threaten to obsolete autologous products before they become profitable?

Lessons Learned

Epicel, acquired by Genzyme via the 1994 merger of its tissue repair business with Biosurface Technologies to form GTR [See Deal], wasn't anyone's concept of an optimal product. At its peak, Genzyme didn't expect it to deliver more than $10 million in sales, mostly for the treatment of third-degree burns.

But a small investment in infrastructure was acceptable to Biosurface/Genzyme, whose strategy was to get that first product into the market to prove the principle of cell-based therapy, then to follow on with larger-market products like Carticel, which was a high-volume product for thousands of patients, not hundreds like Epicel.

While Carticelfar more than Epicel demanded centralized manufacturing and therefore increased turnaround time, the production economics otherwise were more favorable: it required less physical space per patient to produce than Epicel; in its final form, Carticel is just a clump of cells, as opposed to Epicel, which has to be grown more precisely (so that the cells stratify appropriately, because if the layer is too thick, the cells won't take and too thin a layer wouldn't protect) and has to be laid down in gauze for delivery. Also, the amount of cells needed to cover a burn patient is much larger than the number of cartilage cells needed for repair.

Indeed, comparing the characteristics of Epiceland Carticel points up many of the problems of living-skin-graft products generally, including allogeneic products. Nonetheless, two allogeneic skin-graft products are entering the commercial phase, backed by large companies. Last year Novartis Pharma AG , a division of Novartis AG , began selling Apligraf, a two-layered artificial skin product composed of fibroblasts, keratinocytes, and collagen for the treatment of venous leg ulcers, under a 1996 global marketing license from Organogenesis Inc. And a joint venture between Advanced Tissue Sciences Inc. and Smith & Nephew PLC [See Deal] is selling Dermagraft, comprising fibroblasts grown on a framework, for the treatment of diabetic foot ulcers, on a limited basis in Europe. (It has also filed a PMA in the US.) Not surprisingly, like GTR, Advanced Tissue is also moving into living-cell cartilage repair with an articular cartilage product for repairing cartilage defects in joints, through a separate joint venture with S&N. [See Deal]

The principal benefit of an allogeneic versus an autologous product intended for a similar use is that it removes the issue of timing from cell extraction to administration. With GTR's Epicel, what was conceived of as a 21-day preparation at first took closer to four weeks from wound incidence/cell extraction to treatment. Preserving sterility while growing the cells was difficult—the cell-culture process didn't always work, often because the antibiotic in which the biopsy was bathed wasn't effective, which meant going back to get another biopsy sample from critically ill patients where time was of the essence. Allogeneic cell expansion may also fail, and it has a number of economic downsides, but those consequences are invisible to the physician and the patient.

Biosurface/Genzyme did learn a few tricks during the commercialization of Epicelthat streamlined production and delivery, however, which aided acceptance. For example, by changing the media in which the cells grew, the company ultimately shortened product preparation time to 14-18 days. For "fast-track" patients, it would even grow half the cells needed and ship them, then go back and grow the second half.

But the product could never be prepared off-site because the tissue culture process was too difficult to transfer, notes Elma Hawkins, PhD, former director of corporate development at Genzyme and now SVP at vaccine developer Antigenics Inc. If its customers had prepared the grafts, GTR would have still been liable for production and quality problems. On the other hand, to save a day or two by preparing Epicel at the treatment site didn't make sense for a niche product where the added cost and logistics of shipping by courier were less significant considerations than they would have been for a large-market product.

Reimbursement also proved to be a significant, but ultimately manageable, hurdle: using Epicelto repair an arm might cost $200,000. And the acute wound-healing setting also carried with it other reimbursement issues: many third-degree burn patients are mentally unstable—their burns were often deliberately self-inflicted or the result of drunken behavior. (One company executive recounts being present when a burn patient leaped out a window and committed suicide shortly after receiving a graft [not of Epicel].)

Proof of Principle Redux

Commercialization of autologous cancer vaccines is now following a similar course. The first products won't necessarily be the best opportunities, but will prime the market for acceptance of the technology and will serve as a testing ground for the inevitable delivery problems—a task made somewhat easier when dealing with cancer vaccines which, whether autologous products or traditional adjuvants, are generally lower-profile products with poor track records and therefore only modest expectations for efficacy. These smaller opportunities allow companies to build slowly, with modest investments—a good thing since both small market sizes and lack of scaleability of autologous production processes gives Big Pharma little incentive to invest in the area.

Autologous cancer vaccines are a form of active immunotherapy, in which a patient is vaccinated with an antigen—either a "pan-tumor" or "shared" antigen identified by assaying across a panel of allogeneic cell lines representing different tumor types, or one or more antigens taken directly from the patient. The goal of active immunotherapy is to trigger or boost a specific immune response against a tumor by better presenting the antigen, which presumably is also expressed on the surface of the patient's tumor cells but is not being recognized as foreign by the immune system. Such vaccines often incorporate dendritic cells, a naturally occurring type of antigen-presenting cell, which have been purified from the patient.

Active immunotherapies may be made of traditional off-the-shelf adjuvants, or they may have one or more autologous components. For example, the antigens may be purified from a tumor biopsy taken from the patient or created from tumor-associated antigens taken from a cell line sourced elsewhere. And whether shared or patient specific, the antigens may then be fused with antigen-presenting cells that have been extracted from that patient's blood.

The extra steps of taking biological material from a patient, purifying it, and expanding a cell culture, all with the requisite quality controls, make autologous cancer therapies similar to their wound-healing cousins, and very different from off-the-shelf therapy where a physician only has to write a prescription for a drug that has been prepared long before the decision to treat is made. As such, they are a dramatic change in the way medicine is practiced, raising the issue of the timeframe for acceptance and adoption.

On the other hand, hospital pathology labs handle and prepare tumor biopsies and are familiar with safety protocols to prevent cross-contamination. And oncologists, unlike many other medical practitioners, are aggressive adopters of novel therapies. Therefore, the amount of priming the oncology system requires, including the infrastructure changes needed to deliver autologous therapy, is somewhat less than that required in the wound healing surgical setting.

But while their interactions with pathology labs may make oncologists more comfortable with the sample collection process, including the waiting time for cells to be processed and expanded, using those labs to actually produce a vaccine on a commercial basis is another matter. For one thing, farming out the production process assumes technology that may not be in place—designing kits so that laboratories can produce vaccines of reproducible quality, as opposed to having hospital or commercial labs send the biological material to a company-run facility. It may not always be in a company's best interests to decentralize manufacturing because of the additional regulatory questions it would create. And because tumors grow at different rates, notes Gail Maderis, president of Genzyme Molecular Oncology , a division of Genzyme Corp., production of autologous vaccines may not be amenable to standardization in kit form at all.

In addition to technology issues, Maderis points out that for that decentralized model to work hospitals must also make a commitment—dedicating space for compartmentalization, segregation, and handling of the products and training personnel to accommodate the production process.

GMO, which is currently in early trials of a dendritic-cell fusion vaccine for breast cancer at Dana-Farber Cancer Institute and has trials ongoing elsewhere of fusion vaccines for melanoma and kidney cancer using shared antigens, recently invested $1 million to build out a GMP facility intended to produce enough material to get it into pivotal studies. Like most other companies, for now it has decided to avoid the issues a decentralized model creates for rolling out product. Dana-Farber currently makes the breast cancer vaccine, but Maderis wants to shift production to a centralized facility as soon as possible. "Dana-Farber is not used to working under SOPs," she notes, which has slowed the progress of the trial.

However, working with clinical sites to build SOPs (standard operating procedures) into early trials is important because at some point, GMO and other companies will have to convince regulatory agencies that it has in place the necessary release criteria and other measures for assuring quality of patient-specific therapy, which includes interactions with sites for sample collection and initial processing. "If the data are strong, we can solve these problems," Maderis says.

"Product control is critical," agrees Antigenics' Elma Hawkins, in support of the centralized, company-controlled product preparation model for autologous cancer vaccines. "You really need a closed system to transfer a technology off-site to a lab, and it will take years of true development to get that to work…Cell samples will be unique, and each patient is unique," she notes. "You have to think patient by patient."

Like GMO, Antigenics also decided to build out a portion of its 30,000 square foot R&D and laboratory facility to manufacture its first product, Oncophage,which consists of heat shock proteins (intracellular peptide "chaperones") purified from a patient's tumor sample fused to the entire repertoire of antigens found on the surface of that individual's tumor cells (see "Can Heat Shock Proteins Improve Immunotherapy?," START-UP, February 1999 [A#1999900020). Indeed, Hawkins argues, taking the lab operation to pilot scale for late-stage clinical trials as production requirements grow is a simple, serial process for autologous vaccines. "I would rather go into development of something modular," than have to create and validate a truly scaleable manufacturing process. "The up-front investment is small compared, for example, to the scale-up required to produce a biological by fermentation. That's a costly, up-front investment that can't be undone, as opposed to a modular system, which can expand according to the need as a product nears commercialization."

Steve Sherwin, MD, president and CEO of gene therapy and cancer vaccine developer Cell Genesys, disagrees. "At some point you have to reach a commercial phase. If we take our patient-specific lung cancer product through to commercialization, we'd want to know that the therapy kits, preps, quality control, etc. were scaleable when we go to the FDA and propose to go on the market," he says. "It's like a beta-test." And if the product were limited to a small number of patients where scale was not an issue, he questions the business opportunity.

Higher Hurdles

Cell Genesys is involved in developing both autologous and off-the-shelf cancer vaccines. "To the extent it's possible to make an off-the shelf product, the practical advantages in commercialization—and even at the development stage—will always make it the preferred route," Sherwin believes. "The issues go beyond commercial infrastructure. It's also the economic realities of reimbursement and the profit margins of an autologous therapy. Data would have to show the superiority of a patient-specific therapy over an off-the-shelf product."

Cell Genesys'vaccine program, GVAX, originated at Somatix Therapy, which Cell Genesys acquired in 1997. [See Deal] The GVAXplatform comprises tumor cells (in most cases sourced from an in vitro cell line, not a patient-specific tumor) that have been genetically modified to secrete an immune-stimulating protein, granulocyte macrophage-colony stimulating factor (GM-CSF), then irradiated for safety. At the time Cell Genesys merged with Somatix, GVAX was being tested in melanoma, which along with renal cell carcinoma, are the most common targets for immunotherapy.

Unlike in other cancers, the host immune system appears to already be engaged in an ongoing battle against melanomas and renal cell carcinomas, Sherwin explains, as evidenced by anecdotal clinical observations of the more frequent, albeit still unusual, examples of spontaneous remissions observed in these tumor types as compared to other cancers. Using an immunotherapy approach to further boost an already active immune response therefore makes particular sense, especially as proof of concept. But Cell Genesys decided to target prostate and lung cancers with GVAX, for two reasons: market considerations and also, according to Sherwin, to set the bar higher, because obtaining clinically significant results in prostate or lung would more strongly suggest that GVAXwas a platform for the treatment of many cancers, and thus would be a broader commercial opportunity.

Lacking consensus within the clinical community about whether a cell line, which may not express the same antigens as an individual patient's tumor, could nonetheless be used to activate the immune system—a debate that has yet to be resolved—Sherwin wanted to test both autologous and off-the-shelf approaches and evaluate the data. Clearly he prefers the latter option, and not merely for commercial reasons. It won't be possible to make a patient-specific vaccine for every cancer because, depending on the tumor type and its location in the body, the tumor cells needed to make the vaccine won't always be available. "We didn't always have good access to tumor biopsy material," Sherwin says.

Nonetheless, Cell Genesys is developing a patient-specific form of GVAXfor lung cancer, in part because it found a way to do so that was both economical and medically valid. In its original formulation, the company used a retroviral vector to insert the GM-CSF gene into the patient's tumor sample. But a retrovirus can only infect dividing cells, and it took weeks of cell culture to formulate the product once a tumor biopsy was taken.

As with GTR's initial experience with Epicel, that was a nonstarter. Processing time is both a cost issue and a psychological hurdle for patients—another reason why autologous therapies must prove superior to off-the-shelf approaches. And sometimes a patient's disease can outrace the production process. In one Phase II trial of another company's autologous cancer vaccine, only one in five patients could be treated because of an inability to gather sufficient material and grow enough cells to make the product before the disease progressed.

Cell Genesys has now switched to an adenoviral vector, which can insert the GM-CSF gene into non-dividing cells, enabling overnight vaccine preparation on-site at the hospital. Genzyme tried to acquire Cell Genesys in part to support the cell therapy program of its Genzyme Molecular Oncology division [See Deal], which is also counting on the use of adenoviral vectors for better, more sustained presentation of shared antigens. (The deal fell apart when the value of Cell Genesys' stake in its spin-off, antibody-maker Abgenix Inc. , soared—pricing Cell Genesys itself out of Genzyme's reach.)

Although Cell Genesys' goal remains the development of off-the-shelf vaccines, the short manufacturing turnaround time for a patient-specific form of GVAXusing an adenoviral vector makes it economically feasible, Sherwin says. Nonetheless, it's still more complicated than an off-the-shelf product, necessitating on-site training in the use of a GVAX "therapy kit" that the company would provide.

Like Cell Genesys before it switched to an adenoviral vector, many of the first-generation cell therapy companies were clearly technology limited. "The issues of developing allogeneic versus autologous products turn on technology as much as the disease you're trying to ameliorate," acknowledges Tom Okarma, PhD, MD, president and CEO of Geron Corp. and former head of cell-therapy pioneer AIS. Indeed, they were so limited that most of the major first-generation non-wound care cell therapy players failed: AIS, Systemix, and CellPro are all defunct, for one reason or another. (CellPro lost a patent battle to Baxter International Inc. and, bankrupted, eventually sold its assets to Nexell Therapeutics Inc. , the inheritor of the Baxter business. [See Deal])

Okarma joined Geron in 1997 based on the belief that Geron's telomerase platform has potential applications to sustain the viability of differentiated cells produced from stem cells—even when stem cells can be coaxed to expand in cell culture, they are senescent, he points out. That could solve the cell expansion problem faced by AIS and other companies, giving him a second chance to positively impact the development of commercially viable autologous therapies. He expects Geron to be an enabler of autologous therapies by packaging telomerase vectors in a kit, which will be sold to the hospitals, or companies, that are creating the cells.

"The potential for cell therapy, in a visionary sense, is the same," argues Okarma. "The question is what's do-able with current technology." In the early '90s, AIS could only think in terms of a hospital-based mode of delivery for its product candidates, which included CD8 tumor-infiltrating lymphocytes extracted using its proprietary cell-separation technology (see "Rhone-Poulenc's Giant Leap,"IN VIVO, January 1995 [A#1995800033). AIS had to go the patient-specific route, he says, because it lacked a uniform source of tissue and ways to scale cell expansion. And when it established regional manufacturing centers for its products because it saw the need to be local, the company discovered that these multiple company-controlled sites were too expensive.

"Transportation issues—getting samples, extracting the lymphocytes, shipping them back—drove us crazy at AIS," Okarma recalls. And these issues were—and still are—tied to the inherent properties of the technology. Nonetheless, whereas the consumables needed for AIS's first-generation technology weren't readily transferrable to a hospital, companies believe they can now successfully use the hospital environment, which has gained experience in complex QC procedures such as blood banking, to deliver autologous therapy. "That's what has to happen for autologous therapies to succeed," Okarma says. "Integrating technology into that environment and adopting a razor/razor blade business model." Moreover, once the value of autologous therapies is recognized, because hospitals view routine procedures as a significant revenue source, he believes they will eagerly embrace cell culture. "Then, the issues of whether one needs to drive change in medical practice become minor."

Different Views on Driving Change

Like Cell Genesys' Steve Sherwin, GMO's Gail Maderis reasons that if autologous products are only equivalent to existing therapies, the medical infrastructure may not be willing to deal with the intricacies of delivery. "If they work well," she says, "the infrastructure will find a way to cope with the complexities."

Although medical centers will jump on the bandwagon if offered a significant business opportunity, she suggests the attraction is also to be on the cutting edge. Take MRIs and the revenue side of radiology: Radiology may be a significant revenue source of a hospital, but the initial sell for MRIs was the opportunity to be state-of-the-art, she points out, "not as a profit opportunity. The economics shake out later." And because comprehensive cancer centers already do a lot of cell processing and manipulation of tumors for diagnostics and transplants, "for them to pick up a cell therapy isn't a stretch."

As a practical matter, the best way to prep the medical establishment for cell therapy is to challenge with "smaller skirmishes, not a major battle, first," opines Annemarie Moseley, MD, PhD, president and CEO of Osiris Therapeutics Inc. , which last year switched from developing autologous cells for wound healing to focus on allogeneic mesenchymal stem cells (MSCs), which it believes will pass the immune system undetected. Moseley, who was formerly director of clinical gene therapy and data management at Systemix and, before that, VP, clinical research, at AIS, suggests that while all cell-based therapies may perturb the practice of medicine somewhat, it's something that must be minimized so that physicians' use of and experience with cell therapies, and the awareness of success, will drive adoption.

Moseley doesn't only worry about physician acceptance: the health care payers present another hurdle. Osiris thus has a two-sided goal: to minimize the impact of the time and added cost of delivery with allogeneic products, and to demonstrate that the extra steps are worth it. "A company can't expect to come in with too radical an approach," she says, "because medical practice is too formulated, too regimented." Providers need to see this additional aspect of therapy as a positive event, she explains. "Change has to be done gently. We have to be conservative with our first product introductions. If you can't get past the perception, you can't start."

While her comments may not apply as readily to the more innovative arena of cancer vaccines, Moseley's focus is the wound healing market, the target for Osiris's MSCs, which the company intends to bottle, cryopreserved, so that they can be grown in a lab for five days. Allogeneic products are already paving the way in wound healing, she points out, including Apligrafand Dermagraft. "With those products as a model, a clinician just has to understand the need for a five-day window to order the product." To go from that to autologous cell therapy requires factoring in the need for cell collection, expansion, and reinsertion. "It's the same kind of planning, but it's also the difference between a little forethought [to plan an allogeneic procedure] and a lot of forethought," she says.

In the long run, however, Moseley agrees that cell-based medicine will significantly change the practice and attitudes of physicians, providers, and consumers. And she also believes that clinical processing labs—which she refers to as "hybrid pharmacies"—are currently ready, willing, and able to do what's necessary (washing, formulating, loading, and centrifuging samples) to make delivery seamless and not burdensome to the surgeon. "Chemotherapy is prepared in a hospital pharmacy," she notes. And blood banks could also be a point for delivery.

Banking on the Blood Banks

That's what Dendreon Corp. is counting on. Rather than fitting its products into the existing medical framework, Dendreon is developing a hub-and-spoke strategy, using a combination of external centers for cell collection and company-run or licensed cell processing centers to produce its dendritic cell vaccines, the most advanced of which, Provenge, for treating prostate cancer, is in Phase III trials.

Dendreon's is a three-step process. First, dendritic cells are extracted from peripheral blood cells using leukapheresis, the process normally used to purify and separate blood products in a blood bank setting. A patient therefore will go to a blood processing facility with which Dendreon has a contractual relationship—the "spokes," presumably close to the patient's home—to undergo a three-hour white blood cell extraction process. The WBCs are then sent to a cell processing center, either directly managed by Dendreon or one with which the company has a relationship (the "hubs"—in addition to its own centers in Mountain View, Calif., and Seattle, Wash., the company currently has such relationships with The Mayo Clinic, the American Red Cross blood processing facility, Progenitor Cell Therapy, a New Jersey contract cell therapy center, and, through a collaboration with Kirin Brewery Co. Ltd. [See Deal], a center in Tokyo). There, operators who have been trained by Dendreon isolate dendritic cell precursors, then culture them with the vaccine and incubate the mixture for 36 hours to generate antigen-loaded dendritic cells. Finally, the patient is infused with the vaccine in a thirty-minute outpatient procedure not unlike the administration of chemotherapy.

Dendreon initially developed the process at its Mountain View facility, but it learned the ins-and-outs of transferring the technology via its Mayo collaboration, which began in 1997. [See Deal] "We can now get a cell culture facility up and running in three months," says EVP and CSO, Dave Urdal, PhD, including erecting a prefabricated soft-wall clean room and meeting air-quality specifications to prevent sample cross-contamination. The centers can be expanded in a modular fashion, and because the collection process is tied in to an existing blood banking infrastructure, Dendreon doesn't see a large bricks-and-mortar issue associated with its cell therapy.

But Doug Armstrong, PhD, CEO of Aastrom Biosciences Inc. , sees a significant difference between merely having the processes to grow a cell type or mixture of cells for particular tissues and having the capability to convert a small-scale, variable cell culture and expansion process to commercial scale. "Our original plan focused on the first part of the business—learning how to grow bone marrow cells using cord blood. But once we had a research-stage process, because cell culture is as much an art as a process, we realized the importance of the second capability—reproducibility."

Technician-to-technician variability is the fundamental issue in commercialization of cell products, he says. To take out that human element, Aastrom determined it had to completely automate the cell culture procedure. The result: its AastromReplicellautomated culture system, a set of single-use cell culture kits with three compartments to manage fluid and gas delivery that it has priced at $4,000-7,000, depending on the cell type to be cultured. Moreover, he suggests the same barrier to commercial success exists for immunotherapy products such as vaccines derived from patient-specific tumors, so Aastrom is building into its automated platform the capability to bring in other cell types, like T cells or dendritic cells, for use with immunotherapy, and also autologous bone and cartilage.

Importantly, Armstrong believes use of Aastrom's technology only reworks current medical practice to the extent that a physician will need to write a prescription for cells, rather than for a drug. But he also notes the importance of properly preparing cells before they are put into its culture system: "If you put the right cells in, you get the right cells out." For its CB-1kit, a site thaws cord blood, checks cell viability, and inoculates the cells into the kit. "Twelve days later," Armstrong says, "you have a bag of cells." And with other kits requiring a different cell prep, QC can be performed on the inoculum after expansion.

Will Labs Take Center Stage?

Even if Aastrom succeeds in reducing the cell culture process to a much simpler and reproducible set of steps, laboratories will still play a significant role in sample preparation. In any event, should autologous therapies prove their value medically, as Geron's Tom Okarma notes, labs will be integral participants, potentially as manufacturing sources.

Indeed, the personalization of medicine is moving labs closer to the center of medical management. And not just on the cell-therapy side. In the case of the genetic-information plays, their involvement is now extending beyond genetic testing to the collection and banking of tissue samples used for pharmacogenomic profiling. The introduction of prognostic, diagnostic, and therapy monitoring tests is also on the horizon as drugs and molecular diagnostics are developed in tandem.

Already a significant brick in the medical infrastructure, they can only increase in importance.

The Nitty-Gritty of Facilities Management

Manufacturing at both Antigenics and Curis Inc. is headed up by former employess of Genzyme Tissue Repair, who can draw on their experience with GTR's Epicel and Carticel. Curis is gearing up for commercialization of its first cell-based product, Chondrogel (cultured autologous chondrocytes in a matrix), now in Phase III trials, which is used as an injectible bulking agent for treating vesicoureteral reflux, a congenital pediatric condition in which urine flows back from the bladder to the ureter and kidneys.

Jim Sigler, VP, manufacturing, had been director, manufacturing, for GTR, where he was responsible for commercial-scale manufacture of Carticeland Epicel, until 1998, when he joined Reprogenesis Inc. which, along with Creative Biomolecules Inc. and Ontogeny Inc., merged to form Curis earlier this year [See Deal]. Carticel production, Sigler points out, because of its higher volume, carried with it a more complicated set of scheduling and materials management considerations than did Epicel—logistics, warehousing, manufacturing controls, and even where to place customer service.

Most of the hang-ups at GTR were in transitional issues, Sigler says. Epicel, for example, is a high-value product hand carried by courier. If GTR needs to send four boxes of product with a total value of $100,000, shelling out for an airplane ticket is a relatively minor consideration. "But a higher-volume product demands reliable but cheaper logistics." For Curis' Chondrogel, that means using Federal Express to ship biopsy kits back and forth. (Requiring a lab to use the company's kit, which is only $10-20 of material, isn't a technology issue as much as a QC issue.) Curis prefers that sites not inventory kits, but request them patient by patient, so that samples don't arrive back at the company unexpectedly. That's a scheduling and work-flow consideration: cartilage tissue can easily sit for a few days before initiating a culture without affecting viability, and Sigler says that once Chondrogelreaches the market and demand is predictable, the company may supply physicians with kits in advance.

Regardless, Curis believes that, at the outset, it should have an agreed-upon schedule with the physician for when treatment may be administered. The rule of thumb is seven weeks from treatment decision to administration—three weeks to harvest cells and grow them in a flask, including QC to assure viability, another three weeks for expansion in roller bottles, and a week for transit, digestion, and formulation.

Not only must companies solve the logistics of sample tracking and QC, as a process moves from the bench to commercial scale, research-oriented personnel must be kept motivated and interested, or replaced with more shift-oriented staff with different employment and career goals. At GTR, during the Epicelscale-up in the early 1990s, decision-making was held tightly because of the need to operate under SOPs, but therefore did not allow for much personal growth. Thus, people were leaving that side of the company.

Sigler recalls that at GTR, much of the production staff came from local colleges and universities. They were "up-and-comers" who saw their jobs as stepping stones for advancement. His solution to keeping the techs motivated? When the ramp-up of Carticelbegan, the company kept with the Epicel practice and put a patient's name on the tissue culture flask as well as a number, putting a human face on the production process. "Technicians knew the product they were formulating would be on a patient the next day," he explains, which went a long way toward sustaining morale.