Newer clinical strategies for combining interferon and cytotoxic agents against solid tumours and hematological malignancies

The role of interferons in the treatment of cancer continues to evolve. Despite limited single agent activity against solid tumours, interferons now appear to have an important role as modulators of the activity of a variety of cytotoxic drugs. Clinical benefits have been observed for combinations of interferons and alkylating agents against low grade lymphomas, interferons and dacarbazine against malignant melanoma, and interferons and 5-fluorouracil against gastrointestinal and genitourinary malignancies. Further progress will depend on a grealer understanding of the biology of the interaction.

T I-IE ULTIMATE UTILITY OF INTERFERONS (IFN) IN CLINICAL cancer therapy remains unclear.Discovered in 1957 by Isaacs and Lindenmann, IFNS have undergone successive roles as antiviral agents, antiproliferative agents and biological response modifiers.Despite a paucity of information regarding their mechanism of action, the observation that WN -a.therapy induced prolonged clinical remission in patients with hairy cell leukemia resulted in great hopes for IFNs as anticancer agents.With the exception of modest single agent activity against Kaposi's sarcoma, melanoma and renal cell carcinoma, the uWity of single agent IFN against solid tumours was never realized (1) .Recently , however, both preclinical and clinical data have suggested that the greatest utility of WNs may be their use in combination with cytotoxic agents.This paper reviews the data supporting such a role for IFNS.

PRECLINICAL DATA
WNs represent a family of molecules that are clearly not interchangeable.Any discussion of tl1e role of WNs employed in combination with cytotoxic drugs must take into account the diversity of tl1is family of glycoproteins as this is important when analyzing WN interactions in a historical perspective.Furthermore.the complexity of IFN actions must be well understood in terms of attempting to define a hypothesis for the interaction of WNS and cytotoxic agents at the cellular level.
Originally classified according to their source as leukocyte, fibroblast and immune IFNS , more recent nomenclature is based on sequencing (2).WN -a. was originally described as leukocyte WN.WN -~ as fibroblast IFN and wwy as in1mune WN.There are at least 18 WN -a.nonallelic genes; together with the IFN-~ genes, these constitute an IFN -a./~ superfamily of genes on chromosome 9.The IFN-y gene is located on chromosome 17.IFN-a.and IFN-~, the type 1 IFNS, and IFN-y, the type 2 WN , share a variety of features .These include: activity mediated via a cell surface receptor; a complex signal transduction mechanism involving preformed cytoplasmic proteins; binding of these proteins to an IFN-sensitive response element on the cell genome; and up-or down -regulation of a variety of IFN-stimulated genes (3).
The actions and effects of type 1 and type 2 IFNs are distinct, however (4) (Table 1).Therefore, it is unlikely that all classes of IFNS will interact.with cytotoxic agents in precisely the same fashion.Type 1 and type 2 lFNS have unique cell surface receptors.Pathways of signal transduction differ between WN -a. and lFN-y, with differential 1.ryosine phosphorylation of various cytoplasmic proteins and possibly different kinases employed for the different ligand specific pathways.Therefore, cytotoxic drugs with strong effects on membranes, such as doxorubicin, may interact.differently with type 1 and type 2 IFNs.Furthermore.the DNA binding site for lFN-a.activated proteins, termed the lFN-stimulaled response element, and the protein complex that conlacls the DNA binding site differ from that for IFN-y, termed lhe gamma-activated site and the gamma-activated factor.While some lFN-stimulated genes are regulated by both WN -a. and lFN-y, others are differentially regulated.Thus, cytotoxic drugs that act against cellular proteins or interact with a variety of cellular proteins may interact differently in cells treated with different classes of lFNS.
In conclusion, it is clear that type 1 and type 2 IFNs are not interchangeable.Recognition of these differences is critical in interpreting the results of studies employing combinations of lFNs and cytotoxic agents (7).Failure to recognize these differences could lead to premature rejection of lFN-cytotoxic combinations without fu lly testing lhe spectrum of available ffNS.

IFNs AS MODULATORS OF CYTOTOXIC AGENTS
Early in vitro models employing murine leukemia cells as tumour targets demonstrated that IFNS were capable of enhancing the cytotoxic effects of cisplatin, melphalan and cyclophosphamide.In the early 1980s, Balkwill and colleagues ( 5) employed a tumour model system of human breast or non-small cell lung tumours explanted to immunodeficient mice and demonstrated synergy in vivo for IFN-a. in combination with doxorubicin, cyclophosphamide or cisplatin.There were several important findings from these studies.First, given

y.aomrcon
Combining interferon and cytotoxic agents that the mice were immunodeficient, it was likely that the IFN effects were mediated at the biochemical level rather than via immune stimulation.Second, studies of IFN effects in the cytochrome p450 system suggested that the IFN effects were not mediated by altering activation or catabolism of the cytotoxic drugs via hepatic p450 cytochromes.Third, several experiments demonstrated the schedule dependency of this interaction, with simultaneous administration of IFN plus cyclophosphamide being more effective than sequential administration.This latter observation has been amplified recently in detailed studies employing a P388 murine leukemia model; the survival benefits were dependent on the Urning of administration of IFN and cyclophosphamide (6).Since publication of these early studies, synergy or additivity of IFN with over 30 different cytotoxic agents in vitro or in vivo has been demonstrated against a variety of tumour model systems.These have been reviewed extensively in previous manuscripts (7).

CLINICAL STRATEGIES FOR COMBINING IFNs AND CYTOTOXIC AGENTS
1\vo major strategies have emerged from studies combining IFNs with cytotoxic agents (Table 2) .In the first group, IFNS and cytotoxic drugs are combined against tumours for which both IFN and the cytotoxic drug have known antilumour activity.In the second group, IFNs have been combined with agents such as 5-fluorouracil, cisplatin or cyclophosphamide for which there is good evidence for synergy in preclinical tumour models.Examples of both strategies are discussed below.

COMBINATIONS IN WHICH BOTH IFN AND A CYTOTOXIC DRUG HAVE INDEPENDENT ACTIVITY
Single agent TFNS are active against a variety of solid tumours, primarily renal cell carcinoma and melanoma.as well as low grade lymphomas and chronic myelocytic leukemia (1).
Thus, it is natural to employ them in combination with single agent cytotoxic drugs or with combination chemotherapy against these malignancies.Renal cell carcinoma: Both IFN and vinblasUne have modest activity.This combination has been employed in a va.iiety of schedules with no improvement observed over what would be expected with either vinblastine or IFN employed separately (7).Melanoma: Both WN and dacarbazine (DTIC) have response rates in the 15 to 20% range against metastatic malignant melanoma (1).Early studies from Australia (8) and Italy (9) failed lo show a benefit for the combination (Table 3).Furthermore.a randomized trial ofIFN, DTIC or the combination was terminated adm inistratively after very early results failed to demonstrate a benefit (10).In a prospective randomized phase 3 trial from South Africa (11) of DTIC versus DTIC plus IFN.l11e objective response rate for the combination (50% versus 19% for DTIC alone) was significantly better, objective responses were observed in viscera as well as soft tissues, and the combination produced a 38% complete response rate.Furthermore, there was a survival advantage for the combination that was statistically significant.Two studies from Scandinavia of IFN plus combinations of cytotoxic drugs have produced conflicting results (Table 3).In contrast, studies of IFN plus cisplatin against melanoma have resulted in response rates that are similar to those observed with IFN alone , despite substantial evidence in preclinical systems for synergy (12).The lack of anUtumour effect may have resulted in part from the difficulty in administering these regimens because of substantial toxicities.Low grade lymphoma: Use of IFNS in combination with alkylating agent-based regimens against low or intermediate grade lymphomas blends strategies employing IFNs with drugs that have activity against a selected tumour and employing IFNS in combinations for which there is evidence for synergy.

Y•DO co
Early studies employing small numbers of patients suggested a benefit in terms of response rate and progression-free survival for the combination of IFN and chlorambucil in patients with follicular lymphomas.!FN as maintenance therapy did not appear to be beneficial.
In 1988 U1e Eastern Cooperative Oncology Group (ECOG) conducted a randomized t1ial comparing cyclophosphamide, vincristine, prednisone and doxorubicin without (COPA) or with (I-COPA) IFN-a (13).Patients were required to have stage 3 to 4 intermediate grade disease or low grade disease with unfavourable features (B symptoms.bulky disease or rapid doubling times).The toxicities were comparable in both arms with a slighUy higher incidence of infection in the COPA group and a higher incidence of fevers and neurological symptoms in the I-COPA group.Treatment with the lF'N-containing regimen resulted in comparable objective and complete response rat.es.but prolonged time to treatment failure.duration of complete response and overall survival.This occurred despite significant reductions in the doses of cyclophosphamide and doxorubicin received by patients in the I-COPA group (76% of plarmed dose versus 96% of plarmed dose among patients receiving COPA) .
Subsequent. trials ofIFNs and alkylating agents in low grade lymphomas are shown in Table 4.While not definitive.these trials tend to show an advantage for the combination therapy.Chronic myelocytic leukemia: Preliminary data from Australia (14) and MD Anderson (15) suggest a higher 20A incidence of cyt.ogeneUc complete responses with the combination of IFN plus hydroxyurea than with either agent alone.This remains to be confim1ed in larger trials.

COMBINATIONS OF IFNs AND CYTOTOXIC AGENTS FOR WHICH THERE IS EVIDENCE FOR SYNERGY IN PRECLINICAL SYSTEMS
IFNS have been demonstrated to be synergistic with over 30 cytotoxic drugs in preclinical systems.Evidence of synergy for combinations of IFN with 5-fluorouracil (5FU).cisplalin (COOP) and doxorubicin, among olliers, is extensive and suggests the potential for clinical benefits (7).The experience with IFN plus COOP was discussed above.There was no suggestion of benefit against.melanoma, and the combination was toxic.Likewise, the combination of !FN plus doxorubicin in phase 1 trials demonstrated excessive myelosuppression, and Lhere were also several instances of acute cardiac toxicities.including myocardial infarction.This combination has not been extensively studied (except. in combination regimens such as I-COPA).The combination of IFN plus 5FU is the best studied employing this slrategy.Fluorouracil plus interferon: Based on in vitro data against.human colon cancer cell lines, early clinical trials of 5FU plus IFN were initiated at. the Albert Einstein College of Medicine (AECOM) in 1988.A phase 1 trial in patients with colorect.alcancer employing 5FU, 750 mg/ m 2 /week plus IFN at. 3 to 18xl0 6 U daily demonstrated that the maximum tolerated dose for !FN was 15 to 18xl0 6 U daily, but with substantial fatigue (16).Furthermore, among the patients treated at the lower doses oflFN, there was a 56% objective response rate with one complete responder, whereas, at the higher dose level, there were no responders.Based on these data and data from a phase 2 study in England that employed high dose IFN and low dose 5FU, and which resulted in a response rate of less than 10% in advanced colorectal cancer, investigators at AECOM initiated a phase 2 trial employing intermediate doses of both IFN and 5FU (17).
A phase 2 trial employing the AECOM regin1en resulted in a response rate of 63%, but.with formidable toxicities, including three toxic deaths (two with watery diarrhea followed by sepsis).A subsequent ECOG trial (18) with the same regimen, but more stringent dose modification criteria, demonstrated the efficacy and tolerability of this regimen.Among eight clinical trials employing the AECOM regimen, the overall response rate has been 38% (Table 5).Preliminary results from recently completed phase 3 trials of 5FU/IFN versus 5FU/ leucovorin, and 5FU/IFN versus 5FU alone (19.20).have yielded confusing results, but suggest that.combination therapy may be less active than earlier reports suggest.Interpretation of trials awaits reporting of the full results.
Trials in patients with esophageal carcinoma (21,22) and in patients with bladder cancer, who had failed a cisplalin-conlaining regimen (23).also demonstrated substantial activity.In the AECOM trial, there were two complete responders, including one patient who was rendered operable and remains free from disease after two years.
Two mechanisms of interaction between 5FU and IFN have been studied.In vitro studies against human colon cancer cell lines and a murine adenocarcinoma cell line have demonstrated synergistic cytoloxicity for U1e combination which is reversed by exogenous Lhyn1idine, suggesting th.al Lhe combination is acting primarily al the level of the 5FU target enzyme, lhymidylate synthase (24).IFN appears capable of reversing a 5FU-induced elevation in lhymidylate synthase levels.possibly acting al the post-transcriptional level.Fur-lhem1ore, IFN enhances 5FU -induced depletion of lhymidine triphosphate pools, and treatment with the combination resulted in an increase in DNA double strand breaks compared with 5FU alone (25).These data remain preliminary.
In clinical trials, IFN treatment resulted in an increase in plasma levels of 5F'U in patients receiving intermediate dose bolus or infusional therapy (26.27).In some studies this was dependent.on the dose of IFN employed, although this was not invariable.Furthermore, the reduced folale, leucovorin, partially reversed ilie pharmacoki.neticbenefits of IFN in some studies.which appeared to result from a decrease in 5FU catabolism (28).

CONCLUSION
Combinalions of IFNS and cytotoxic agents have resulted in improvements in objective response rates, Lime to failure and overall survival in various clinical trials.In the case of IFN plus 5FU.mechanisms of inleraclion have been described in detail.The relative importance of each remains lo be determined.Further studies of IFN-cytoloxic combinations are warranted.

TABLE 5 Phase 2 clinical trials with the combination of intermedi- ate dose 5-fluorouracil and recombinant interferon-a*
INFECT DIS VOL 5 SUPPL A FEBRUARY 1994 CAN J