5-AZA-CdR is an effective epigenetic drug for the treatment of hematologic malignancies. The clinical efficacy of 5-AZA-CdR is due to its demethylating epigenetic action, which reactivates tumor suppressor genes silenced by DNA methylation. Low-dose 5-AZA-CdR (20 mg/m2 one-hour infusion x five days or 15 mg/m2 four-hour infusion q8 h × three days) can produce CRs in patients with MDS and AML [99, 123]. These low dose schedules cause less toxicity than intensive doses of 5-AZA-CdR. This is especially important for older patients with a poor performance status who are not good candidates for intensive therapy with 5-AZA-CdR. The interesting responses to low-dose 5-AZA-CdR suggest that leukemic stem cells are very sensitive to low concentrations of 5-AZA-CdR, an indication of the immense chemotherapeutic potential of this drug.
Recent preclinical reports indicate that very low dose 5-AZA-CdR administered frequently (two to three times/week) also has the potential to be an effective form of therapy for cancer. One reason for the failure of cytotoxic chemotherapy is that malignant cells can be resistant to the induction of apoptosis due to a non-functional p53 pathway as a result of mutations or deletions . However, the genes that program terminal differentiation in these apoptosis-resistant malignant cells can be silenced by epigenetic mechanisms, such as DNA methylation, and reactivated by non-toxic doses of 5-AZA-CdR. This same treatment maintains the self-renewal of normal hematopoietic stem cells by preventing repression of stem cell genes by differentiation-inducing stimulus and induces differentiation of AML cells . For these reasons the very low dose 5-AZA-CdR does not produce pronounced granulocytopenia as observed with intense doses of this agent.
Laboratory studies on AML cells support the use of very low-dose 5-AZA-CdR. 5-AZA-CdR inhibits in vitro proliferation, decreases colony formation and induces myeloid differentiation of p53-null AML cells . These observations were confirmed using fresh AML cells from a patient. The AML cells were transplanted into NSG immunosuppressed mice and treated with a s.c. injection of 5-AZA-CdR (0.2 mg/kg three times/week for two weeks, then once/week). This very low dose 5-AZA-CdR was much more effective in prolongation of the survival time of the leukemic mice than an intense dose of ARA-C (75 mg/kg per day intraperitoneally for five days). The proof of principle of the very low-dose 5-AZA-CdR was also confirmed in a clinical trial in MDS patients with high-risk cytogenetics . 5-AZA-CdR 3.5 to 7 mg/m2 administered one to three times/week produced an overall response of 84% (CR + hematologic improvement + stable disease), which is remarkable. Complete cytogenetic remissions were observed in 50% of the patients. It will be interesting to see if this non-toxic differentiation therapy with 5-AZA-CdR will be effective in older AML patients who are not candidates for cytotoxic chemotherapy.
One limitation of the low dose 5-AZA-CdR for the treatment of AML or MDS is the problem of eradicating malignant cells in the liver or spleen due to the high activity of cytidine deaminase. Deamination of 5-AZA-CdR can reduce its concentration to sub-therapeutic levels in these organs. The use of an inhibitor of cytidine deaminase, such as tetrahydrouridine, in combination with 5-AZA-CdR has the potential to overcome this problem . The proof of principle of this approach was demonstrated in a murine xenotransplant model of AML where tetrahydrouridine produced a marked enhancement of the antineoplastic activity of 5-AZA-CdR . The combination of these agents merits a high priority for clinical investigation in patients with hematologic malignancies.
Preclinical studies indicate that very low-dose 5-AZA-CdR also has the potential to be an effective treatment for tumors with a favorable epigenetic signature. As an example, human renal carcinoma cells derived from a patient were inoculated s.c. into nude mice followed by treatment with low dose 5-AZA-CdR (0.2 mg/kg s.c. × 3/week) . This low dose therapy was very effective in reducing tumor growth and did not produce leukopenia. This low dose 5-AZA-CdR, in combination with tetrahydrouridine to inhibit cytidine deaminase, was also very effective in inhibiting the growth of murine melanoma tumors in mice . The low dose chemotherapy did not produce leukopenia or reduction in body weight. 5-AZA-CdR was also shown to induce differentiation of both human and murine melanoma cell lines.
These observations on very low 5-AZA-CdR therapy of leukemia and tumors were confirmed by Tsai et al.  using a different dose-schedule. These investigators showed that low dose 5-AZA-CdR (72-hour exposure) reduced colony formation of AML cells from patients, but not the normal hematopoietic stem cells colony-forming units-granulocyte macrophage (CFU-GM). The low dose 5-AZA-CdR (0.1 μM, 72 hours ex vivo) followed by 7 to 14 days drug-free media was also shown to decrease tumorigenicity in mouse tumor xenografts.
In summary, the very low dose 5-AZA-CdR preclinical studies showed that this type of treatment could produce a loss in the self-renewal potential of cancer stem cells due to the increase in the expression of genes that suppress malignancy. These epigenetic changes are maintained in the target cells after drug removal and accumulate with each low dose treatment until there is a complete loss of cancer stem cell potential. The low dose chemotherapy merits clinical investigation in patients with cancer. The very low dose 5-AZA-CdR may also have the potential to maintain CR in patients with leukemia and arrest malignant progression in patients with solid tumors. For cancer patients with poor performance status the very low dose 5-AZA-CdR therapy may be a good option to improve the quality of life rather than the use of only supportive therapy or no treatment.
Some patients with cancer may not respond or show disease progression on the low dose-schedule 5-AZA-CdR. This may be due to fact that: a) the cancer may have an epigenetic/genetic signature that is not predisposed to the induction of terminal differentiation by low dose 5-AZA-CdR; b) the target cancer cells may have a low level of deoxycytidine kinase, the enzyme that activates the prodrug, 5-AZA-CdR ; c) the cancer cells may be in anatomic sanctuaries that have low penetration of 5-AZA-CdR (for example, CSF, testis, tumors with a limited blood supply ). The concentration of 5-AZA-CdR in these sanctuaries is too low to eliminate the cancer stem cells; d) the cancer cells may be in a biochemical sanctuary that contains high levels of cytidine deaminase (for example, liver, spleen); e) drug resistance develops more rapidly after repetitive treatments with low-dose chemotherapy; f) because 5-AZA-CdR is a cell cycle-specific agent, a one- to four-hour infusion of this agent only targets cancer cells in S-phase, whereas cells in G1 and G2 phases escape the chemotherapeutic action of this analog during short-term treatment. A long interval (12 to 24 hours) between infusions can also permit leukemic stem cells to pass through the S-phase cell cycle without exposure to 5-AZA-CdR and, thus, escape its antileukemic action. This possibility was demonstrated in a preclinical study on leukemia using the S phase–specific drug ARA-C .
One approach to overcome these caveats is to use intensive chemotherapy with 5-AZA-CdR administered as a continuous infusion for patients with leukemia. This objective is of high priority and involves determination of the optimal plasma level of 5-AZA-CdR and duration of treatment that can eliminate leukemic stem cells in these sanctuaries. The in vitro data on colony assays of human leukemic cell lines indicate that a concentration of 5-AZA-CdR in the range of 1 to 2 μM for the duration of the cell cycle of the leukemic cells has the ability to completely eliminate their proliferative potential. Another approach is to use 5-AZA-CdR in combination with an inhibitor of cytidine deaminase, such as tetrahydrouridine or zebularine, to target the leukemic cells in the biochemical sanctuaries [135, 136]. An additional approach is to use 5-AZA-CdR in combination with a biochemical modulator, such as 3-deazauridine, to eliminate drug-resistant leukemic cells due to a deficiency in deoxycytidine kinase . Cancer cells deficient in deoxycytidine kinase are very sensitive to the cytotoxic action of 3-deazauridine .
One of the key points concerning intensive-dose therapy with 5-AZA-CdR is the fact that it produces delayed and prolonged myelosuppression. Several investigators have used intensive-dose 5-AZA-CdR in patients with advanced leukemia and observed that most of these patients with a good performance status recovered from the hematopoietic toxicity. These early studies were performed before the clinical use of granulocyte colony stimulating factors to accelerate the recovery from myelosuppression. Some examples of the dose schedules of intensive-dose 5-AZA-CdR that were used are: total dose of 2,479 mg/m2 as a 60-hour i.v. infusion; 500 mg/m2 as a 6-hour infusion every 12 hours × 5 days for a total dose of 5,000 mg/m2; and a daily dose of 300 to 500 mg/m2 as a 24 to 120 hour infusion [97, 108, 137]. The estimated plasma level of 5-AZA-CdR in these studies ranged from 1 to 3 μM. The hematopoietic toxicity produced by intensive doses of 5-Aza-CdR can also be predicted from its comparative pharmacology with the related deoxycytidine analogue, ARA-C. Both 5-AZA-CdR and ARA-C are S-phase–specific agents. They have identical metabolism, and their antineoplastic action is due to their incorporation into DNA, but their molecular mechanisms of action differ. 5-AZA-CdR inhibits DNA methylation, whereas ARA-C potently inhibits DNA replication. Because they target the same cells (proliferating cells in S phase), they should produce a similar pattern of hematopoietic toxicity. Most leukemic patients in CR with a good performance status recover from the hematopoietic toxicity produced by very high-dose ARA-C (up to 6,000 mg/m2/day for four days; total dose 24,000 mg/m2) . These observations provide a rationale for intensive doses of 5-AZA-CdR in the range of 1,000 mg/m2/day in leukemic patients in CR without encountering unacceptable hematologic toxicity for patients with a good performance status. The recovery from granulocytopenia after 5-AZA-CdR is approximately two weeks longer than that after ARA-C . This is probably due to the delayed epigenetic action of 5-AZA-CdR on normal hematopoietic stem cells compared with the acute cell kill produced by ARA-C. From this point of view, it is better to use a six-week interval between cycles of 5-AZA-CdR rather than the four-week interval used for ARA-C.
It is a remarkable achievement that current chemotherapy can induce CR in most patients with hematologic malignancies. The major challenge is maintaining the patients in CR. Patients in CR are good candidates for experimental chemotherapy because of their good hematologic status. From an ethical point of view, high-risk leukemic patients with an unfavorable karyotype that predicts a poor outcome are good candidates for intensive therapy with 5-AZA-CdR.
Pharmacokinetic/pharmacodynamic calculations can be used to estimate the optimal dose for 5-AZA-CdR. For the initial studies, we recommend a combination of intensive and low-dose 5-Aza-CdR to treat high-risk patients with leukemia. For the initial intensive phase, 5-AZA-CdR can be infused at a rate of 30 mg/m2/hour for days one and two (total dose 1,440 mg/m2/day). This infusion rate should give a plasma concentration of approximately 2 μM, shown to be very effective in both in vitro and in vivo animal studies on leukemia. The objective of this intensive therapy of 5-AZA-CdR is to target the most rapidly proliferating leukemic stem cells and those in anatomic and biochemical sanctuaries. The intensive phase is followed by a low-dose phase where 5-AZA-CdR is administered on days three and four as a short infusion at a dose of 30 mg/m2/day. The objective of this phase is to target cancer stem cells with a long cell cycle that do not enter the S phase during the first two days of treatment. Supportive care with granulocyte colony stimulating factor is recommended to shorten the duration of granulocytopenia. The interval between each cycle of therapy should be six weeks to permit adequate recovery from bone marrow toxicity. Modifications of the proposed dose schedule may be required for optimization. In subsequent studies, the low-dose phase of 5-AZA-CdR can be replaced by histone deacetylase (HDAC) and/or histone methylation inhibitors, which showed a synergistic interaction against leukemic cells in preclinical studies [139, 140]. It could also be interesting to design a clinical study in AML patients to see the efficacy and safety of decitabine in combination with promising novel tyrosine kinase inhibitors such as quizartinib . Both preclinical and clinical observations indicate that 5-AZA-CdR has tremendous potential for the treatment of hematologic malignancies. The results of this proposed clinical trial on 5-AZA-CdR will be of great interest and will hopefully lead to improved overall survival of patients with advanced leukemia.
Chemotherapy of solid tumors using 5-AZA-CdR also merits clinical investigation. Most malignancies have a large number of tumor suppressor genes that are silenced by aberrant DNA methylation, providing many interesting targets for 5-AZA-CdR therapy. The preclinical data indicate that all types of tumors are sensitive to 5-AZA-CdR treatment, including low dose therapy. However, the clinical responses to low-dose 5AZA-CdR in solid tumors are reported to be very limited. It should be noted, however, that the estimated plasma levels of 5-AZA-CdR in these trials was too low. In vitro clonogenic assays using an exposure time of 24 hours on tumor cells indicate that low concentrations of 5-AZA-CdR are not very effective. Our preclinical data on chemotherapy of tumors in the mouse model suggest that the plasma level of 5-AZA-CdR should be approximately 3 μM for curative therapy .
An example of the potential of intensive doses for tumor therapy is the pilot stage III/IV NSCLC study where the patients were administered 5-AZA-CdR (660 mg/m2 as an eight-hour infusion), which produced a plasma concentration in the range of 3 μM. This study produced some promising results: three patients who survived for 15 months and one patient who survived for seven years . The preclinical data support the possibility of this type of response and predict that an infusion time longer than eight hours should be more effective. One major reason for the failure of tumor chemotherapy is the limited penetration of drugs into tumors . One approach to overcome this problem is to obtain high plasma concentrations of anticancer drugs to enhance their penetration into tumors. This provides a rationale for the use of intensive-dose 5-AZA-CdR for the treatment of solid tumors. S phase–specific drugs can be used at very high doses for a limited duration without unacceptable side effects. Patients with metastatic malignancy and poor prognosis are potentially good candidates for intense chemotherapy with 5-AZA-CdR. As an initial study, we suggest a dose schedule of 60 mg/m2/hour administered as an 18-hour infusion (total dose of 1,080 mg/m2). This dose schedule will give an estimated steady state plasma concentration in the range of 4 μM. The interval between cycles should be six weeks to permit adequate bone marrow recovery. Patients treated previously with intensive cytotoxic chemotherapy are at risk of severe hematopoietic toxicity and require a minimum of four weeks of recovery before being eligible for intensive 5-AZA-CdR therapy. Patients who do not respond to anticancer agents that do not produce hematopoietic toxicity would also be good candidates for this investigational therapy. Depending on the response, the intensive dose schedules for 5-AZA-CdR may have to be modified for optimization (for example, increase the dose and/or duration of the infusion; use low dose 5-AZA-CdR between cycles of intensive doses). The goal should be to optimize the dose-schedule of 5-AZA-CdR to reveal its potential for tumor therapy. It may also be possible to increase the effectiveness of this tumor therapy by using a sequential treatment of 5-AZA-CdR followed by an inhibitor of histone modification [139, 140] or with a tyrosine kinase inhibitor .