As researchers gain a better understanding of the unique aspects of individual tumor types and their surrounding microenvironment, the design of novel therapies categorized as prodrugs is become increasingly sophisticated, and several novel constructs show particular promise.
For decades, investigators have been looking for ways to combine the optimal aspects of different types of anticancer therapy to create more effective, safer drugs. As researchers gain a better understanding of the unique aspects of individual tumor types and their surrounding microenvironment, the design of novel therapies categorized as prodrugs is become increasingly sophisticated, and several novel constructs show particular promise.
The term prodrugs refers to a broad group of therapies designed to remain inactive until they reach their intended target. It also encompasses a wide variety of mechanisms for the delivery and activation of anticancer therapies.
Thus far, antibody—drug conjugates (ADCs) have been the most promising class of prodrugs in anticancer therapy, with several agents now approved by the FDA. Beyond that, other prodrugs with diverse mechanisms of action have gained oncology indications.1 A recent article in Nature Reviews Drug Discovery identified at least 30 prodrugs that the agency has approved in the past decade across all clinical categories, including 5 for cancer, that in total account for more than 12% of the small molecule new chemical entities authorized during that time.1
Despite disappointments in the field, some observers expect prodrug strategies to be an important facet of future drug development for cancer and other disease states.1,2The term prodrug was coined in the 1950s3 and has been explored as a drug design strategy ever since. It refers to an inactive or less active derivative of a drug molecule that can be transformed into the active form via some sort of chemical or enzymatic activity.
Prodrugs are designed to modify the physical and chemical characteristics of a drug to make it more effective. A prodrug strategy can help ensure that a drug is active just at a particular time or place to both maximize cancer cell killing and minimize off-target toxicity. It can improve the bioavailability, solubility in water, metabolism, absorption, or route of administration of a drug, as well as its ability to overcome physiological barriers, such as the blood—brain barrier and the gastrointestinal tract epithelium.
The plethora of prodrug strategies that have been tested can be separated into 2 major categories, based on the way in which they are converted into the active form. Passively activated prodrugs exploit the differences between cancerous and normal cells, such as the overexpression of certain cell surface receptors by cancer cells, or the unique nature of the microenvironment surrounding the tumor, such as lower pH, reduced oxygen levels, and aberrant vasculature. Prodrugs can also be “turned on” via active strategies, such as by simultaneously introducing a secondary activating drug or substance to the patient.4-6ADCs represent the most promising form of prodrug therapy to date. They are composed of a monoclonal antibody that targets a specific tumor-associated antigen linked to a chemotherapeutic agent, which, in theory, enables targeted delivery of the cytotoxic “payload” to the tumor (Figure).
The design of the linker between antibody and drug is a vitally important consideration. It must be broken at precisely the right time to ensure maximum efficacy and prevent release of the cytotoxic drug into the circulation. Numerous designs, including acid-sensitive linkers, protease-sensitive linkers, and glutathione-sensitive linkers, are intended to exploit the low pH of the tumor microenvironment, the specific enzymes expressed by tumor cells, and the high levels of glutathione, respectively.6-8
The FDA has approved 4 ADCs for these targets and tumor types: ado-trastuzumab emtansine (Kadcyla) for HER2 in metastatic breast cancer; brentuximab vedotin (Adcetris), CD30 in Hodgkin and anaplastic large cell lymphoma; inotuzumab ozogamicin (Besponsa), CD22 in acute lymphoblastic leukemia (ALL); and gemtuzumab ozogamicin (Mylotarg), CD33 in acute myeloid leukemia (AML). A variety of ADCs are among the prodrugs in clinical development, many of which are in phase III trials (Table), suggesting the list of FDA-approved agents may expand in the near future.9-10 \ADCs are an example of a carrier-linked prodrug, wherein the parent drug itself is not modified but attached to a secondary molecule that dictates its activity. Alternatively, prodrugs can involve specific alterations to the parent drug.
Various types of chemotherapy-based prodrugs fall into this second category. Cyclophosphamide is a prodrug of phosphoramide mustard, a DNA cross-linking agent, that is activated only in cells that have low levels of aldehyde dehydrogenase. Temozolomide, another prodrug, is a derivative of the alkylating agent dacarbazine.
Capecitabine, used in the treatment of breast and colorectal cancers, is a prodrug that has little cytotoxic activity itself until it is enzymatically metabolized into the active drug 5-fluorouracil (5-FU), which inhibits DNA synthesis.
A particularly successful prodrug strategy for chemotherapy, however, employs nanoparticle technology. Chemotherapy drugs are essentially stored inside nanoparticles, which improves their solubility and bioavailability, helping more of the drug reach the tumor before it is broken down. The drug is released once cancer cells take up the nanoparticles.11,12
The nanoparticles can also be coated with proteins such as albumin, which help to further increase the specificity of drug delivery to the tumor. This type of prodrug strategy takes advantage of the enhanced permeability and retention effect—the preferential accumulation of small molecules in the tumor microenvironment, which is thought to result from the abnormal “leaky” vasculature of tumors and lack of a lymph drainage system, among other factors.13
Nab-paclitaxel (Abraxane) and aldoxorubicin represent 2 examples. Both are albumin-bound formulations of the chemotherapeutic drugs paclitaxel and doxorubicin, respectively. The FDA approved nab-paclitaxel, which also uses nanoparticles, for the treatment of advanced breast cancer in 2005, for the first-line treatment of advanced non— small cell lung cancer (NSCLC) in 2012, and for treating late-stage pancreatic cancer in 2013.14 Aldoxorubicin was a promising candidate in soft-tissue sarcoma, but plans to submit the drug for FDA approval in this indication have not materialized.15,16 The drug remains in early-stage clinical development as part of immunotherapy combinatons in solid malignancies, including pancreatic cancer (NCT03387098) and triple-negative breast cancer (NCT03387085).
Another drug in development is CRLX101, a nanoparticle formulation of camptothecin conjugated to a sugar molecule, cyclodextrin; it is undergoing phase II clinical testing in patients with metastatic castration-resistant prostate cancer (CRPC). However, disappointingly, recently published results from a phase II study of CRLX101 in combination with bevacizumab (Avastin) demonstrated no improvement in median progression-free survival (mPFS) in patients with metastatic, refractory renal cell carcinoma. Among 111 patients randomized to receive the combination or investigator’s choice of any approved regimen not prevously received, mPFS was 3.7 months versus 3.9 months, respectively.17 Minimal activity was also observed in patients with chemotherapy-refractory gastroesophageal cancer when used as monotherapy.18
In a similar manner, chemotherapeutic drugs can be encapsulated in liposomes or conjugated to natural polymers, such as polyethylene glycol. MM-310 is a liposomal formulation of docetaxel in which the liposomes are conjugated to antibody fragments directed at the ephrin receptor A2 (EphA2), which is overexpressed in a variety of cancer types and the stromal cells of the microenvironment. Upon intravenous infusion, the liposomes target cells overexpressing EphA2; they are internalized, and docetaxel is released.19Another prodrug of 5-FU, flucytosine (5-FC), showed little promise in the treatment of cancer until recently, when investigators started employing it in a different way, highlighting another promising prodrug strategy. Directed enzyme prodrug therapy (DEPT) uses an enzyme that is artificially introduced into the body to activate the prodrug. Several forms of DEPT that differ in how the enzyme is introduced to the cancer cell have been explored. The most promising introduces suicide genes into the cancer cells via gene therapy (GDEPT).
Two main types of enzymes are used in GDEPT: herpes simplex virus thymidine kinase (HSV-TK) and cytosine deaminase (CD). HSV-TK is used in conjunction with guanosine- based prodrugs that were originally developed as antiviral therapies. It converts these prodrugs into toxic triphosphates that cause cell death primarily by blocking DNA synthesis. CD is used in conjunction with 5-FC, which it converts into 5-FU.20,21
Tocagen, a drugmaker based in San Diego, California, has developed Toca FC, a novel extended-release formulation of 5-FC. Investigators are testing this in combination with Toca 511, a retrovirus that encodes the CD enzyme. The phase III Toca 5 trial is comparing Toca 511/Toca FC with standard of care in patients with recurrent high-grade glioma. Tocagen recently reported that enrollment in this trial is complete.22
In a phase I trial, investigators evaluated this therapy in 56 patients with high-grade glioma who recurred after standard of care therapy and underwent surgical resection. Toca 511 was injected into the resection cavity wall, and Toca FC was subsequently administered orally in ascending doses; 23 patients received the dose being evaluated in the pivotal phase III trial.
Among all patients, the objective response rate was 11.3%. In a post hoc analysis of the patients who received phase III dosing, the ORR was 21.7%, with 5 complete responses (CRs). Complete and partial responses began between 6 and 19 months after Toca 511 administration, suggesting an immune-based response. Median duration of response had not been reached after a median follow-up of almost 3 years.23
TG6002 is a modified oncoloytic Vaccinia virus, in which the genes for thymidine kinase and ribonucleotide reductase have been deleted to restrict propagation of the virus to tumor cells; it expresses a bifunctional suicide gene FCU1 that is a fusion of the CD and UPRT genes, encoding a chimeric enzyme.24
The virus preferentially infects tumor cells and produces the FCU1 gene, which encodes for the FCU1 enzyme. When 5-FC is administered alongside the virus, it is broken down in the FCU1-expressing tumor cells into 5-FU. A phase I/II trial in patients with recurrent glioblastoma is ongoing at Groupe Hospitalier Pitié-Salpêtrière in Paris, France (NCT03294486).
ProstAtak, meanwhile, uses the HSV-TK system. An adenoviral vector (AdV-TK) containing the gene that encodes HSV-TK is injected into a tumor as the prodrug valacyclovir, an antiviral drug, is administered simultaneously. A protein in the vector phosphorylates the prodrug, leading to cell death and the release of tumor neoantigens. This sets off a chain of signaling via the STING pathway that ultimately results in the activation of T cells that promote cytotoxic immunotherapy. ProstAtak is being evaluated in the phase III ULYSSES (NCT02768363) and PrTK03 (NCT01436968) trials in patients with prostate cancer.25
In addition to its cytotoxic effects, this type of prodrug therapy is thought to have a dual mechanism of action, with additional antitumor efficacy resulting from the promotion of an immune response. Areas of low oxygen concentration, or hypoxia, are a characteristic feature of many tumors thought to result from the disorganized, aberrant, and hyperpermeable network of blood vessels that surrounds them. Hypoxia has been shown to contribute to more aggressive and treatment-resistant tumors and poorer patient prognosis.
Hypoxic regions can also present significant challenges for effective treatment. They tend to contain cells that grow and divide more slowly, which is problematic for the use of chemotherapy that targets rapidly dividing cancer cells. Furthermore, these regions may be beyond the reach of the normal vasculature, which can hinder drug delivery.
This has prompted the development of hypoxia- activated prodrugs (HAPs), designed to metabolize into active drugs in a hypoxic environment. HAPs are designed to become active following biological reduction, consisting of chemotherapeutic agents conjugated to a bioreductive protecting group. This ensures that they are activated only by specific enzymes that chemically reduce and break down the protective group and release the active drug. The enzymes that catalyze this reaction are oxidoreductases.26,27
The prime example is evofosfamide (TH-302). It consists of a bromo-isophosphoramide mustard cytotoxin conjugated to 2-nitroimidazole, which acts as the protective group. The 2-nitroimidazole group is reduced by oxidoreductases, generating an oxygen-sensitive radical ion. When oxygen is plentiful, the radical ion is converted back into the prodrug, but in hypoxic conditions, it is further reduced and broken down to release the active drug.
Evofosfamide demonstrated promising response rates in phase II clinical trials for the treatment of advanced pancreatic cancer and soft-tissue sarcoma.28,29 However, phase III trials failed to show an overall survival (OS) advantage.30,31 Clinical development continues for both evofosfamide and several analogs.
Apaziquone is another HAP in development, but it is designed to function as such only in cells that have low expression of a particular oxidoreductase enzyme, NAD(P) H:quinone oxidoreductase 1 (NQO1), a 2-electron oxidoreductase. Both 1- and 2-electron oxidoreductases are involved in bioreductive processes; however, only 1-electron oxidoreductases are oxygen sensitive. The presence of endogenous 2-electron oxidoreductases has been a confounding factor in the development of HAPs because they can reduce prodrugs even in the presence of oxygen.27,32,33
Apaziquone initally showed promise in treating patients with nonmuscle invasive bladder cancer when administered via intravesical instillation after transurethral resection of the bladder tumor (TURBT). Unfortunately, the FDA declined to approve it based on findings from 2 phase III trials—both missed their primary endpoint of a statistically significant reduction in disease recurrence at 2 years.34
A recently published pooled analysis of the trial, which was not included in the initial study designs, suggested that the timing of apaziquone relative to surgery may be important. Efficacy was increased when apaziquone was instilled within 60 minutes of TURBT.35 Ongoing phase III trials are testing this, in addition to the effects of administering a second dose of apaziquone (NCT03224182, NCT02563561).
Because HAPs rely on the expression of oxidoreductases to facilitate their effects, the expression levels of these enzymes may prove to be an important biomarker of sensitivity to these drugs.Several small molecule—based prodrugs are also in development. OBI-3424 targets cancers overexpressing the aldo-keto reductase 1C3 enzyme and releases a DNA alkylating cytotoxic drug in the presence of this enzyme, the expression of which has been observed in a number of cancer types, including T-cell ALL.
OBI-3424 was granted an orphan drug designation in September for this indication and, a few months prior to that, for the treatment of patients with hepatocellular carcinoma (HCC).36 Enrollment is under way for a phase I/II trial in patients with HCC, CRPC, and other solid tumors (NCT03592264).
Another member of this class is guadecitabine (SGI-110), a small molecule prodrug of the DNA methyltransferase inhibitor (DNMT) decitabine. It consists of decitabine linked to deoxyguanosine, which shields the active drug from degradation in the intracellular compartment.
Guadecitabine has shown the most promise in patients with AML; however, the results of the phase III ASTRAL-1 trial—in which 815 patients were treated with guadecitabine or physician’s choice of azacitidine, cytarabine, or decitabine—did not demonstrate an improvement in the coprimary endpoints of CR rate and OS.37 Numerous clinical trials of guadecitabine are ongoing. Prodrugs of another DNMT inhibitor, 5-azacitidine (NUC013), are in preclinical development.38,39