FEATURE

Therapeutic cancer vaccines on trial

Therapeutic cancer vaccines offer hope for patients for whom traditional treatments have failed,
but various obstacles may impede the launch of these new agents.
Janice M. Reichert and Cherie Paquette
2002 Nature Publishing Group http://biotech.nature.com

Cancer is the second-leading cause of death

in the United States, responsible for one in
four fatalities and contributing to $180 bil-
lion in direct medical costs. The American
Cancer Society estimates that approxi-
mately 1.3 million new cases of cancer were
diagnosed in the United States during
2001, with seven of the most common can-
cers accounting for 68% of these cases
(Table 1). Today, a diagnosis of cancer is
not necessarily a death sentence: there were
nearly 9 million cancer survivors living in
the United States in 1999 (ref. 1), suggest-

Tom & Dee Ann McCarthy, Corbis

ing that current cancer treatments are
effective. However, the standard options
surgery, radiotherapy, and chemothera-
pyhave debilitating and distressing side
effects, destroying healthy tissues along
with rogue cancer cells.
Fortunately, biotechnology has opened
up avenues for the development of gentler
and more selective treatments. For exam-
ple, the Food and Drug Administration Will vaccines become an integral part of the treatment of cancer patients in the future?
(FDA; Rockville, MD) recently approved
two new antibody-based chemothera-
pies: IDEC Pharmaceuticals (San Diego, could stop cancers growing in the first immune system by secreting immunosup-
CA) Zevalin, a monoclonal antibody instanceprophylactic cancer vaccines pressive agents and expressing molecules
attached to the radioisotope Yttrium-90, will be further investigated. Then, at-risk that can induce the death of white blood
and Wyeth Pharmaceuticals (Collegeville, patients could be inoculated against the cells2,3. However, regardless of whether the
PA) Mylotarg, a monoclonal antibody cancers that threaten them the most. cancer is invisible to the immune system,
linked to the anti-tumor agent calicheam- To date, research has yielded several can- or is simply not immunogenic enough,
icin. Unlike standard chemotherapies, cer vaccines that are now on the market, cancer vaccines are designed to boost an
these monoclonal-based agents are and over 50 products that are in clinical tri- anticancer immune response.
designed to deliver their toxic payload als. This retrospective study presents an In theory, the mode of action of a cancer
specifically to cancer cells, leaving healthy overview of the clinical development of vaccine is simple: the vaccine prompts the
cells unharmed. therapeutic cancer vaccines, highlights the immune system to produce anti-tumor
Nevertheless, all too frequently, patients challenges to the development of these antibodies and cytotoxic T lymphocytes
face a relapse during treatment or a recur- products, and finally questions whether (killer T cells), which target, kill, and clear
rence of their cancers after treatment is these challenges will become obstacles
complete. Currently, basic and clinical to getting these potentially life-saving
research is centered on developing so- products on the market. Table 1. New cases of cancer diagnosed in the
called therapeutic cancer vaccines for United States during 2001
patients who already have cancer. Once the Natural born killers Type of cancer Total number of new
safety and efficacy of the current genera- Cancers develop because the bodys cases (both sexes)
tion of therapeutic products has been surveillance system fails. Exactly why Prostate 198,100
established, next-generation vaccines that the immune system fails to detect and Breast 193,700
eliminate cancer cells is still poorly Lung 169,500
understood, but several factors likely Colon/rectum 135,400
Janice M. Reichert is senior research fellow contribute. First, cancers arise from Non-Hodgkins lymphoma 56,200
and Cherie Paquette is research assistant at the malfunction of healthy cells, so Bladder 54,300
Melanoma 51,400
the Tufts Center for the Study of Drug they may not elicit an immune
Development, Boston, MA response because they are regarded by Source: American Cancer Society, Cancer Facts & Fig-
(janice.reichert@tufts.edu and the body as self. In addition, tumors ures 2001 (http://www.cancer.org/downloads/STT/
F&F2001.pdf).
cherie.paquette@tufts.edu). can actively curb the function of the

http://biotech.nature.com JULY 2002 VOLUME 20 nature biotechnology 659

 FEATURE

malignant cells2. In real- 15

ity, achieving the desired Vaccines (n = 69)
immune response is not Companies (n = 32)
so simple: the cancer
vaccine must stimulate 12
the immune system in
such a manner that
healthy cells are spared
Number of trials
9
and only cancerous cells
destroyed. To achieve
2002 Nature Publishing Group http://biotech.nature.com

this end, cancer vaccines

currently under clinical 6
study incorporate anti-
gens (foreign molecules
that stimulate the hosts

Amy Center
immune system) specif- 3
ic to a particular cancer
type. However, cancer
antigens alone are weak-
0
ly immunogenic, and 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001
the use of additional
immunostimulator y Year of first clinical study
molecules is generally
necessary. Furthermore, Figure 1 Clinical trials of cancer vaccines initiated annually between 1985 and 2001. Year of first clinical study of cancer
autoimmunity against vaccine was unknown for one product and one company.
native antigens on
healthy cells might be induced if the anti- bryonic antigen (CEA) is the target for sev- senting cells, which includes monocytes,
gen is not sufficiently cancer specific4. eral colorectal cancer vaccines, whereas macrophages, and antibody-secreting B
Therefore, cancer vaccines must produce a MUC-1 and HER-2 are target antigens for lymphocytes (B cells), and they can be cul-
delicate balance between two undesirable several breast cancer vaccines. tured in large numbers in vitro7. The DCs
states of the immune systemunderstim- Whole cells displaying cancer-associated are loaded with the desired antigen using
ulation and overstimulation. antigens can also be used as vaccines. Cells electroporation and can also be genetically
can be derived from two sources: the can- modified to secrete an additional immune-
Many means to an end cer itself and the immune system. In the response stimulant such as granulocyte
Cancer vaccines consisting of antigens of first instance, cancer cells are killed (usual- macrophage colony stimulating factor
varied composition, identity, and source ly by irradiation), then modified either (GM-CSF).
have been studied clinically. Products genetically or chemically to increase their Whether made up of isolated antigens or
might consist of antigens that are recombi- immunogenic potential. The identity of the whole cells, cancer vaccines can be derived
nant proteins, synthetic peptides, carbohy- tumor-specific antigens need not be from either the patients own cells (autolo-
drates, extracted tumor-derived proteins, known because the cell itself becomes the gous products) or from those of human
or monoclonal antibodies. Alternatively, vaccine and the immunological key to the donors (allogenic products). The specifici-
the product might be DNA encoding the destruction of its cancerous relatives. The ty of autologous preparations might
antigen of interest. The identity of the anti- second method involves producing den- increase the effectiveness of the treatment,
gen used depends on the type of cancer, dritic cells (DCs) that directly present the but the manufacture of such vaccines is
although some antigens are associated with tumor antigen to the immune system. DCs labor intensive and time consuming, and
multiple types5,6. For example, carcinoem- are one of a group of so-called antigen-pre- products would not be available immedi-
ately. On the other hand, allogenic prod-
ucts may be easier to produce, being made
from commonly expressed antigens (such
Analysis criteria as CEA), and would be available off the
The biopharmaceuticals database of the Tufts Center for the Study of Drug Development shelf ; allogenic products are not, however,
includes data on 70 cancer vaccines sponsored in clinical studies by 33 US-based patient specific and therefore could be less
companies. Initiation of clinical studies for the products occurred between January 1985 and effective.
December 2001. Data were collected by surveys of the sponsoring companies and from
public documents. Of the 70 cancer vaccines, 52 were in clinical development and 18 Cancer vaccines in the clinic
products were discontinued. The year the sponsoring company initiated its first cancer The pharmaceutical and biotechnology
vaccine clinical study could be assigned for 32 of the 33 companies. The year clinical study industrys interest in cancer vaccines is
was initiated could be assigned for 69 of the 70 products. If the sponsoring company in- revealed here in an analysis of the number
licensed the product, the first clinical study might have been done by an academic or of companies that initiated studies of these
government (e.g., National Cancer Institute) investigator. Data for BCG products marketed in products during the past decade (Fig. 1,
the United States as treatments for bladder cancer were not included in analyses because and see Analysis criteria). From 1993 to
similar products were previously approved as prophylactic vaccines. These products are 2001, 28 companies that had not previously
marketed by non-US-based companies (Table 4). JR, CP developed cancer vaccines initiated clinical
studies of at least one product, and at least

660 nature biotechnology VOLUME 20 JULY 2002 http://biotech.nature.com

 FEATURE

three products have entered clinical study NCEs approved in the United States from Hodgkins lymphoma. One cancer vaccine
each year since 1993 (Fig. 1). However, few 1996 to 1998 were in clinical development is in clinical studies as a treatment for
companies have expanded their develop- for an average of 81 months9. In compari- small-cell lung cancer.
ment programs for cancer vaccines during son, cancer vaccines in phase 3 have already As the identities of potential antigens are
this period. The majority (61%) of compa- been in clinical development for a mini- established, competing companies might
nies sponsored clinical studies of just one mum average of 64 months (Table 3), and develop vaccines based on the same target
product, and only five (15%) sponsored phase 3 will probably take longer than but composed of different molecules (such
clinical studies of more than three cancer phase 2 (38.5 months). These data suggest as peptides, proteins, or DNA) and involv-
vaccines. Development of cancer vaccines that the clinical development of the cancer ing the use of different immunostimulato-
might be more aggressively pursued when vaccines likely to be considered for ry adjuvants. For example, at least five
2002 Nature Publishing Group http://biotech.nature.com

more products are approved and confi- approval earliest will take a minimum melanoma vaccines incorporating the
dence and experience in developing this average of 102 monthsan even more gp100 antigen have entered clinical studies
type of product has increased. protracted timeline than that for the sponsored by four different companies.
On the other hand, companies seem to clinical development of anticancer mAbs Indeed, the wide variety in the type of
have maintained their commitment to the and NCEs. cancer vaccines in phase 3 studies might
products that they moved into clinical tri- reflect the lack of consensus on the best
als. To date, only 18 of the 70 products that way to produce an immune response to
entered clinical study have been discontin- Indeed, the wide variety in the cancerous cells. No single antigen type
ued (Table 2). In total, 14 products are in makes up a majority of the productssix
phase 1, 24 products are in phase 2, and 14 type of cancer vaccines in are either peptide or protein antigens, four
products are in phase 3. The products that phase 3 studies might reflect are whole cells (tumor cells or dendritic
entered clinical study the earliest have cells), two are monoclonal antibodies, one
advanced in development the furthest. Of the lack of consensus on the is antigen-encoding DNA, and one is a gan-
the products that entered clinical study best way to produce an immune glioside antigen (Table 4).
during 1985 to 1989, four (67%) are in
phase 3. Nearly half of the products that response to cancerous cells. Marketed products
entered clinical study during 1990 to 1994 Three cancer vaccines developed by US-
are in either phase 3 (26%) or phase 2 based companies are currently marketed
(16%). Of products that entered clinical Previous studies have already shown that outside the United States (Table 4):
study during 1995 to 1999, only four (11%) clinical development of biopharmaceuti- OncoVax is marketed in The Netherlands
are in phase 3, whereas none of the cancer cals in general is taking longer10. There are by Intracel (Frederick, MD) as a treatment
vaccines that entered clinical study during many possible reasons why clinical devel- for colon cancer; Avax Technologies
2000 to 2001 are beyond phase 2. opment is being slowed: studies have (Overland Park, KS) M-Vax is licensed for
So far, the pace of clinical development become increasingly complex because the treatment of melanoma in Australia;
for cancer vaccines is similar to that of more procedures are required for each and Corixa (Seattle, WA) markets Melacine
monoclonal antibody (mAb) anticancer study participant (for example, if a specific in Canada for patients with stage IV
therapeutics. Approximately half of the 77 side effect is identified, companies feel melanoma.
mAbs currently in development or under obliged to check that parameter in all Both OncoVax and M-Vax are modified
review by the FDA are being studied as future studies, resulting in an add-on effect autologous tumor cells, and because the
treatments for cancer8. The anticancer for screens) and study costs are rising. In cells come from different patients, they
mAbs currently in phase 2 were in phase 1 addition, mergers and acquisitions and the have variable composition and inherent
for 17 months, whereas those currently in in-licensing of products may delay clinical batch-to-batch variation. The products are
phase 3 were studied for an average of 19 development programs. therefore not regulated as therapeutics in
months in phase 1 and 41 months in phase The Netherlands or Australia but are cus-
2. By comparison, cancer vaccines current- The vaccine landscape tom made in processing facilities that are
ly in phase 2 were studied an average of 20 Fourteen cancer vaccines are currently in licensed by the health authorities in the
months in phase 1, whereas those currently phase 3 of clinical development in the two countries. Melacine is an allogenic
in phase 3 were studied an average of 25 United States (Table 4), and these products cancer vaccine consisting of lysed cells
months in phase 1 and 38.5 months in are being studied as treatments for five of from two human melanoma cell lines com-
phase 2 (Table 3). Phase 3 length, typically the most prevalent cancers in the United bined with a proprietary adjuvant. The
the longest, cannot be calculated for the States. Four are potential treatments for product contains at least eleven
cancer vaccines because none of the prod- melanoma, and four are being studied as melanoma-associated antigens (HMW-
ucts in the analysis cohort have completed treatments for
phase 3. cancers of the
The duration of clinical trials (combin- digestive sys- Table 2. Current status of cancer vaccines in the United States
ing phases 13) for anticancer mAbs and tem (colon, designated by year of first entry into clinical trials
new chemical entities (NCEs) approved in rectum, stom- 19851989 19901994 19951999 20002001 Total
the United States might provide a bench- ach, and pan-
mark for the likely duration of clinical creas). Two Phase 1 0 0 8 6 14
Phase 2 1 3 17 3 24
development of cancer vaccines. Currently, products are Phase 3 4 5 4 0 14*
five anticancer mAbs are approved in the being studied Discontinued 1 11 6 0 18
United States. These mAbs were in clinical for prostate Total number 6 19 35 9 70
development prior to first approval for an cancer, and
average of 83 months. The 11 anticancer two for non- *Year of first clinical study unknown for one product in phase 3.

http://biotech.nature.com JULY 2002 VOLUME 20 nature biotechnology 661

 FEATURE

Potential obstacles ahead?

Table 3. Mean and median phase lengths for cancer vaccines and anticancer mAbs Cancer might become more common in
Therapeutic product Phase 1 (months) Phase 2 (months) the United States, the European Union, and
Japan because the average age of their
Cancer vaccines in phase 2 (n = 24) 20.1 (20.4) x
Cancer vaccines in phase 3 (n = 14) 25.3 (20.4) 38.5 (41.0) populations is increasing. In the United
Anticancer mAbs in phase 2 (n = 17) 17.0 (17.5) x States, the number of cancer patients over
Anticancer mAbs in phase 3 (n = 8) 19.1 (20.2) 41.1 (47.4) the age of 85 is expected to increase four-
fold between 2000 and 2050 (ref. 1).
Median values are given in parentheses. mAbs, monoclonal antibodies; phase 1, interval from start date of
first phase 1 study to start date of the first phase 2 study; phase 2, interval from start date of the first phase 2
Progress in medicine already extends the
lives of many cancer patients, and so the
2002 Nature Publishing Group http://biotech.nature.com

study to start date of first phase 3 study; x, the value cannot be calculated for products currently in this phase
of development. number of people experiencing relapse or
developing metastases may well increase in
the future. The therapeutic cancer vaccines
MAA, Melan-A/MART-1, gp100, TRP-1, S- bladder. The BCG therapeutics are different currently in clinical development have the
100, GD2, GD3, MAGE-1, MAGE-2, substrains of attenuated Mycobacterium potential to help this population of cancer
MAGE-3, and tyrosinase). bovis, which first entered clinical studies in patients.
All three marketed products are currently 1921 as a vaccine against tuberculosis11. The However, there are several obstacles that
in clinical trials in the United States. first marketed therapeutic cancer vaccine could prevent new products from reaching
However, the FDAs recent requests for infor- was, therefore, derived from a traditional the market. First is the enormous number
mation about the composition and charac- prophylactic vaccine. The BCG cancer vac- of tumor antigens, adjuvants, and strate-
terization of two other cancer vaccines cines are injected directly into the bladder, gies that could be used to create therapeu-
Dendreons (Seattle, WA) Provenge and and locally stimulate the immune system to tic cancer vaccines: it will not be possible to
CancerVaxs (Carlsbad, CA) Canvaxin kill the cancerous cells. Unlike the antigen- study every permutation in patients to find
suggest that the FDA has yet to clarify the based vaccines currently in development, the optimal product, so effective methods
regulation of these cell-based products. the exact mode of action of the BCG prod- for weeding out preclinical candidate vac-
The only cancer vaccines marketed in the ucts is unknown. Therefore, the BCG prod- cines are needed. Second is the inexperi-
United States are three bacillus ucts are of limited use as models for the ence of the sponsoring companies, the
CalmetteGurin (BCG) products, which cancer vaccines currently in clinical devel- majority of which are biotechnology com-
are treatments for carcinoma in situ of the opment. panies, each with fewer than three cancer

Table 4. Therapeutic cancer vaccines in phase 3 and beyond in the United States
Company Trade name Vaccine type* Indication(s) Year first marketed,
(code name) country availability

Intracel OncoVax Autologous cancer cells Colon cancer 2000, The Netherlands
Titan Pharmaceuticals CeaVac Anti-idiotype mAb Colorectal cancer NA
AVI BioPharma Avicine Peptide antigen Colorectal cancer NA
Corixa Melacine Allogenic tumor antigens Melanoma 1999, Canada
Vical Allovectin-7 Gene for antigen Melanoma NA
CancerVax Canvaxin, CancerVax Allogenic cancer cells Melanoma NA
Progenics Pharmaceuticals (GMK) Ganglioside antigen Melanoma NA
Genitope (Idiotype immunotherapy) Recombinant protein antigen, Non-Hodgkins NA
patient specific lymphoma
BioVest International (Idiotype protein vaccine) Autologous protein antigen Non-Hodgkins lymphoma NA
Aphton Gastrimmune Recombinant protein antigen Pancreatic, colorectal, NA
stomach cancer
ImClone BEC2 Anti-idiotype mAb Small-cell lung cancer NA
Dendreon Provenge Modified autologous DC Prostate cancer NA
Northwest Biotherapeutics DCVax-Prostate Modified autologous DC Prostate cancer NA
Antigenics Oncophage Autologous protein antigen Renal cell cancer, NA
melanoma
Avax Technologies M-Vaxa Autologous cancer cells Melanoma 2000; Australiac
Shire Pharmaceuticals PACIS BCGb; live bacteria Bladder cancer 1994; Canada,
(United Kingdom) Philippines, Argentina,
and United States
Aventis (France) TheraCys, ImmunoCys BCGb; live bacteria Bladder cancer 1990; Canada,
Germany, United States,
Argentina, Turkey,
Singapore, Australia,
and Hong Kong
Akzo Nobel (The Netherlands) TICE BCG BCGb; live bacteria Bladder cancer 1989; worldwide

*mAb, monoclonal antibody; DC, dendritic cell.

aM-Vax clinical program was inactivated during 2001; AVAX anticipates filing an Investigational New Drug application for M-Vax produced using modified

production process later in 2002.

bBCG, bacillus CalmetteGurin.
cAustralian Medical Services Advisory Committee is currently reviewing patient reimbursement.

NA, product not on the market

662 nature biotechnology VOLUME 20 JULY 2002 http://biotech.nature.com

 FEATURE

vaccines in clinical development and no Act of 1997), designed for therapeutics that Acknowledgments
marketed therapeutics. The pharmaceuti- have the potential to address unmet med- Funding for this work was provided in part by a
cal industry can offer little guidance: the ical needs and can treat serious and life- grant from the Drug Information Association (Fort
few multinational pharmaceutical compa- threatening diseases. The program provides Washington, PA).
nies that specialize in vaccines develop pro- sponsors with additional avenues for com-
phylactic rather than therapeutic vaccines, munication with the FDA during the clini- 1. Edwards, B.K. et al. Annual report to the nation on
and these generally protect against infec- cal development of a product with a Fast the status of cancer, 19731999, featuring implica-
tions of age and aging on US cancer burden.
tious diseases. Third is the novelty of the Track designation. Currently, only 2 of the Cancer 94, 27662792 (2002).
products: the cancer vaccines are disparate 52 cancer vaccines in development have this 2. Armstrong, A.C. et al. Science, medicine, and the
enough that the experiences of one compa- designation. In addition, these products future: Cellular immunotherapy for cancer. BMJ 323,
2002 Nature Publishing Group http://biotech.nature.com

12891293 (2001).
ny might not be useful to another. might be eligible for priority review, reduc- 3. Dalgleish, A.D. Cancer vaccines. Br. J. Cancer 82,
The most pressing issue is whether these ing the time until the FDAs first action (the 16191624 (2000).
4. Armstrong, A.C. & Hawkins, R.E. Vaccines in oncol-
obstacles will result in delays in approvals first formal response to the application sub- ogy: background and clinical potential. Br. J. Radiol.
for the products. Certainly, the novelty of mitted) on the licensing application to six 74, 9911002 (2001).
the products could be a drawback during months or less. 5. Durrant, L.G. Cancer vaccines. Anticancer Drugs 8,
727733 (1997).
the approval process because of the relative Therapeutic cancer vaccines have the 6. Pardoll, D.M. Cancer vaccines. Nat. Med. 4,
lack of experience and scientific expertise potential to change the standard of care for 525531 (1998).
7. Timmerman, J.M. & Levy, R. Dendritic cell vaccines
at the FDA in reviewing cancer vaccines. cancer patients. However, the companies for cancer immunotherapy. Annu. Rev. Med. 50,
Potentially, there could also be a higher developing these products face a number of 507529 (1999).
level of scrutiny for therapeutic cancer vac- obstacles, in part because of the diversity 8. Reichert, J.M. Therapeutic monoclonal antibodies:
trends in development and approval in the US. Curr.
cines, which act by less well-studied mech- and novelty of the products. The challenge Opin. Mol. Ther. 4, 110118 (2002).
anisms than those of anticancer mAbs and for the future will be to rationally design 9. Kaitin, K.I. & Healy, E.M. The new drug approvals of
1996, 1997, and 1998: drug development trends in
NCEs. Similar factors might also hinder products based on optimized strategies, the user fee era. Drug Inf. J. 34, 114 (2000).
approvals in the European Union and target only the most promising candidates 10. Reichert, J.M. et al. The effects of the Prescription
Japan. emerging from preclinical models, and Drug User Fee Act and the Food and Drug
Administration Modernization Act on the develop-
In the United States, companies can carefully design clinical studies in relevant ment and approval of therapeutic medicines. Drug
speed the process of clinical development patient populations to ensure that cancer Inf. J. 35, 8594 (2001).
11. Chung, K.-T. & Biggers, C.J. Albert Lon Charles
by using the FDAs Fast Track program (leg- vaccines reach the patients that so desper- Calmette (18631933) and the antituberculous BCG
islated as part of the FDA Modernization ately need them. vaccination. Perspect. Biol. Med. 44, 379389 (2001).

http://biotech.nature.com JULY 2002 VOLUME 20 nature biotechnology 663