Combination therapy with 2-difluoromethylornithine and a polyamine transport inhibitor against murine squamous cell carcinoma



Using a recently developed autochthonous mouse model of squamous cell carcinoma (SCC), a combination therapy targeting polyamine metabolism was evaluated. The therapy combined 2-difluoromethylornithine (DFMO), an inhibitor of ornithine decarboxylase (ODC), and MQT 1426, a polyamine transport inhibitor. In 1 trial lasting 4 weeks, combination therapy with 0.5% DFMO (orally, in the drinking water) and MQT 1426 (50 mg/kg i.p., bid) was significantly more effective than with either single agent alone when complete tumor response was the endpoint. In the combination group, 72% of SCCs responded completely vs. 21 and 0% for DFMO and MQT 1426, respectively. A second trial involved a 4-week treatment period followed by 6 weeks off-treatment. With apparent cures as an endpoint, combination therapy was again more effective than either agent alone: a 50% apparent cure rate was observed in the combination group vs. 7.7% in the DFMO group. MQT 1426 had no inhibitory effect on SCC ODC activity nor did it enhance the inhibition by DFMO, but SCC polyamine levels declined more rapidly when treated with combination therapy vs. DFMO alone. The apoptotic index in SCCs was transiently increased by combination therapy but not by DFMO alone. Thus, targeting both polyamine biosynthesis and polyamine transport from the tumor microenvironment enhances the efficacy of polyamine-based therapy in this mouse model of SCC. © 2005 Wiley-Liss, Inc.

In 2004 ∼28,300 people in the US were diagnosed with squamous cell carcinoma (SCC) of the head and neck (oral cavity and pharynx) and ∼7,200 died of the disease.1 The management of advanced SCC of the head and neck, especially recurrent tumors following resection, is currently very difficult. Available chemotherapeutic approaches for recurrent disease rarely lead to favorable outcomes. New therapies based on knowledge of signaling pathways important to cancer progression have recently proven to be successful against various forms of cancer.2, 3, 4, 5 One cellular pathway long known to be deregulated in many epithelial cancers, including SCC, is polyamine metabolism.6 An important gatekeeper enzyme in this pathway is ornithine decarboxylase (ODC), which converts L-ornithine to the diamine putrescine. Subsequent biosynthetic enzymes convert putrescine to spermidine and spermine via successive transfer of propylamino groups derived from S-adenosylmethionine. Early attempts to exploit polyamine metabolism using the ODC inhibitor 2-difluoromethylornithine (DFMO) in humans were largely unsuccessful,7 although there are no published reports of treatment of SCC. However, recently DFMO has shown efficacy in mouse models of SCC. In a mouse model of UV-induced skin tumorigenesis, DFMO at 0.4% in the drinking water caused regression of squamous tumors (including SCCs) although the effect was transient.8 We have recently reported effective therapy of SCC with DFMO in a newly developed autochthonous mouse model of SCC, in which tumor progression is driven by an ODC transgene.9 In this model, tumors develop rapidly and in high multiplicity (up to 4 SCCs/mouse) after a single treatment with carcinogen, making it an efficient model for production of SCCs.

A potential factor limiting the therapeutic efficacy of DFMO is the ability of tumor cells to take up polyamines from their microenvironment via a specific membrane transport system.10 The specific sources of imported polyamines could be the tumor blood supply or, especially during therapeutic regimens, extracellular polyamines released by dying tumor cells. Both sources could, in theory, be used to replete cellular polyamine levels in tumor cells undergoing regimens involving DFMO. With the recent availability of a high affinity inhibitor of polyamine transport,11 we reasoned that combination therapy of SCCs with both DFMO and a polyamine transport inhibitor (PTI) might improve the therapeutic response to DFMO. Combination therapy might also permit the use of lower doses of DFMO than when typically used in therapy or prevention studies (1–3%, orally). The results reported here suggest that combination therapy is a promising approach for the treatment of SCC.

Material and methods


The structure of the polyamine transport inhibitor (PTI) MQT 1426 is illustrated in Figure 1. MQT 1426 was synthesized using an adaptation of the method previously described for the synthesis of L-Lys-spm (ORI 1202).11, 12 The synthesis can be summarized as follows. The 4-nitrophenyl ester of α,ε-Boc-D-lysine was coupled to the free base of spermine in methanol. The remaining amine groups of the conjugate were then protected as their Boc-carbamates and the resulting protected molecule was purified by SiO2 chromatography. The Boc-carbamates were hydrolyzed using methanolic hydrochloric acid to give the pentahydrochloride salt of MQT 1426. This molecule was characterized by 1H and 13C NMR, LC/MS, HRMS and elemental analysis and its structure conformed to all spectral data obtained. DFMO as a hydrochloride salt was generously supplied by ILEX Oncology, Inc. (San Antonio, TX), now Genzyme Oncology (Cambridge, MA).

Figure 1.

Structure of MQT 1426 (D-Lys-spermine).

Animals and treatments

The K6/ODC mouse model on the FVB/N strain background instead of the original C57BL/6J background13 was used in all experiments. All experiments with mice were approved by the Institutional Animal Care and Use Committee of Lankenau Institute for Medical Research. To induce SCCs, newborn (1 day old) pups were treated once topically with 100 nmol of DMBA dissolved in 50 μl of acetone. Both male and female mice were used. SCCs typically developed on the treated area (dorsal skin) beginning 5 weeks later. SCC-bearing mice were used in experiments between 8 and 18 weeks of age. Based on pilot data indicating a general lack of responsiveness of SCCs of the face and head, these tumors were excluded from analyses.

There was substantial variability in tumor growth rates, even among multiple SCCs in the same mouse. In an attempt to control for this variability, tumor-bearing mice were randomized based on tumor size distribution to 1 of 4 treatment groups: control (phosphate-buffered saline; vehicle for MQT 1426), DFMO (either 0.25% or 0.5% in drinking water), MQT 1426 and DFMO plus MQT 1426. Thus, each treatment group consisted of mice bearing SCCs with approximately equal tumor size distributions (small→large). DFMO solutions were changed every 4 days or less. MQT 1426 was dissolved in phosphate-buffered saline at a concentration of 9.75 mM (5 mg/ml) and injected i.p. at a dose of 50 mg/kg twice daily, in early morning and late afternoon, except on weekends when mice were injected once. Animals were observed daily for signs of distress due to tumor burden and drug treatment. Tumor volume was calculated according to the equation V = l × w2/2 where l = length (longest dimension) and w = width. Responses of individual tumors to treatment were expressed as “ratio of tumor volume,” which is defined as the tumor volume after treatment divided by the initial (pretreatment) volume.

Tumor harvest, ODC and polyamine measurements

SCC-bearing mice were rapidly euthanized, SCCs excised and a piece or pieces placed in Fekete's fixture for histopathologic diagnosis and the remaining tumor was placed immediately on dry ice and transferred to −80°C for subsequent biochemical determinations. ODC activity and polyamine levels were measured in buffer extracts and 0.2 N perchloric acid extracts of tissue, respectively, as previously described.14, 15 Buffer extracts for ODC determination from DFMO-treated mice were dialyzed overnight vs. a 1,000-fold excess of buffer to remove free DFMO. One unit of ODC activity corresponds to 1 nmol of CO2 liberated/hr. Polyamine levels are expressed as nmol/mg DNA.

Measurement of apoptosis and cell proliferation in SCCs.

Tumor-bearing mice were injected i.p. with BrdUrd (100 μg/g body weight) 1 hr before sacrifice. Pieces of tumor were fixed overnight in Fekete's solution and processed for paraffin embedding. Assessment of apoptotic and cell proliferation indices was performed using a TUNEL-based assay and detection of incorporation of BrdUrd into nuclei, respectively, as previously described.8


Characteristics of MQT 1426

In a drug development program to identify potent PTIs, a series of amino acid–spermine conjugates was synthesized.12 We selected MQT 1426 (D-lysine-spermine; Fig. 1) for our studies because of its desirable PTI properties: using previously published methods,10 the Ki for inhibition of spermidine uptake in MDA-MB-231 cells was determined to be 28 ± 11 nM (n = 3) and the EC50 for growth inhibition in the same cell line was 3.1 ± 1.5 μM (n = 3). Acute toxicity studies with i.p. delivery in mice indicated an LD50 of ∼150 mg/kg/day. A pharmacokinetic study demonstrated rapid drug clearance from plasma (t½ = 10 min) following i.v. administration of 26 mg/kg. Thus, we chose a total daily dose of 100 mg/kg/day (i.p.) administered twice a day ∼8 hr apart for our experiments. This dose was well tolerated as no mice died during treatment, nor was significant weight loss observed.

Efficacy of DFMO plus MQT 1426 against murine SCCs

Our previous study of DFMO as a single agent against SCC in the K6/ODC.FVB model employed high doses (1–2%, given orally). The objective of combination therapy with DFMO and a PTI was to use a lower dose of DFMO, still retaining a therapeutic effect. The results of our initial trial of the combination of the polyamine synthesis inhibitor DFMO and the PTI MQT 1426 on the growth of SCCs are shown in Figure 2 and Table I. The dose of DFMO chosen for this trial was 0.5%. Compared to control SCCs, tumors in mice given either DFMO alone or MQT 1426 alone grew more slowly (Fig. 2). The mean ratios of tumor volume for both the DFMO alone and MQT alone groups was greater than 1.0, yet significantly lower then the control group (p = 0.04 and p = 0.02, respectively). However, the combination of DFMO and MQT 1426 was significantly more effective than either agent alone, with a mean ratio of tumor volume less than 1.0. More importantly, after a 4-week treatment period, there were many more complete responses in the combination group than in either of the other treatment groups (Table I). In the combination group there were 13 complete responses out of 18 tumors treated. The only other group with complete responses was the DFMO group, in which 3 of 13 tumors completely regressed. The difference in complete response rate between these 2 groups is highly significant (p = 0.011; Fisher's Exact Test). MQT 1426 treatment alone produced some partial responses, indicating that at least a subset of SCCs is dependent on the activity of the polyamine transport system for growth. Some characteristics of these MQT 1426-responsive SCCs are described below (see “Effect of DFMO with or without MQT 1426 on ODC activity and polyamine levels”).

Figure 2.

Mean tumor response by treatment group after 4 weeks of treatment. SCC-bearing mice were randomized based on tumor size distribution to 1 of 4 treatment groups: control (vehicle for MQT), 0.5% DFMO alone, MQT at 50 mg/kg bid i.p. and the combination of DFMO plus MQT. Data points represent the ratio of tumor volume, defined as tumor volume at 4 weeks after treatment divided by pretreatment tumor volume. For graphical purposes, 12 tumors in the DFMO and MQT group were plotted with values as 0.01, although the actual value for each was 0.0. p = 0.04 (DFMO vs. control); p = 0.021 (MQT vs. control), p = 0.005 (combination vs. control) and p = 0.004 (combination vs. DFMO) (all 2-sided t tests.)

Table I. Summary of Tumor Responses in Study 1
  1. Complete response indicates a 99% or greater volume reduction, partial response is a tumor volume smaller after treatment than the pretreatment tumor volume. No response was scored if the tumor increased in volume during the treatment period.

MQT 14261301408
DFMO + MQT 142618132102

A second combination therapy trial with a slightly different experimental design was conducted to further evaluate the complete responses observed in the first trial. In this study, the DFMO dose was lowered to 0.25% and after the 4-week treatment period, we observed the mice for a 6-week off-treatment period so as to detect possible cures. The results of this trial are shown in Figure 3 and Table II. With this experimental design, SCCs in all active treatment groups showed accelerated tumor growth after an initial response to treatment. The “rebound” effect was most prominent in the DFMO alone and MQT alone groups; less so in the combination therapy group. Thus, in the combination group, during the 6-week off-treatment period, 3 of 10 tumors regrew after cessation of treatment, compared to 10 of 13 in DFMO alone and 9 of 12 in MQT 1426 alone groups (Table II). Although the 6-week off-treatment period may be too short to draw definitive conclusions, there were 5 apparent cures in the combination therapy group and 1 in the DFMO alone group (p = 0.05 for DFMO plus MQT 1426 group vs. DFMO alone; Fisher's Exact Test).

Figure 3.

Mean tumor response by treatment after 10 weeks (4 weeks of treatment, 6 weeks of follow-up). SCC-bearing mice were randomized based on tumor size distributions to 1 of 4 treatment groups: control (vehicle for MQT 1426), 0.25% DFMO, MQT 142 at 50 mg/kg bid i.p. and the combination of DFMO and MQT 1426. For graphical purposes 5 tumors in the DFMO and MQT group and 1 tumor in the DFMO group were plotted with values of 0.01 although the actual values were 0.0. There were no significant differences in mean ratio of tumor volumes among treatment groups (p values > 0.18, pairwise 2-sided t tests).

Table II. Summary of Tumor Responses in Study 2
  1. A complete response indicates a 99% or greater volume reduction at the end of the 4-week treatment period, while cure is defined as the lack of visible tumor at the end of the 6-week off-treatment period. Partial response is defined as a ratio of tumor volume ≤ 1.0 after the 6-week off-treatment period. Regrowth occurred if the tumor increased in volume during the off-treatment period.

10 03100
MQT 142641203  
10 0390
10 01111
DFMO + MQT 142641055  
10 0235

Effect of DFMO with or without MQT 1426 on ODC activity and polyamine levels

To assess the effect of the different treatments on the targeted cellular pathway, ODC activities and polyamine pools were measured in SCCs after short term treatment with DFMO (0.5%), MQT 1426 and DFMO plus MQT 1426. Control SCCs expressed very high levels of ODC, albeit with a substantial variability (Fig. 4). As early as 24 hr after treatment and persisting throughout the 8-day period, ODC activity was suppressed by >90% in both treatment groups administered DFMO (Fig. 4). MQT 1426 at some treatment points appeared to reduce SCC ODC levels but the effect did not reach statistical significance.

Figure 4.

Effect of drug treatments on SCC ODC activity. SCC-bearing mice were randomized based on tumor size distribution, as described in Material and Methods, to 1 of the 4 treatment groups: control, 0.5% DFMO, MQT 1426 (50 mg/kg/day bid) and the combination of DFMO and MQT 1426. At the indicated time after treatment, mice were rapidly euthanized and pieces of SCCs processed for ODC activity. Results are the mean ± SE from 3–5 SCCs per treatment group. *p < 0.01 (vs. control, Students t test).

In contrast to normal tissues, in which putrescine is the least abundant polyamine, in SCCs putrescine accounts for as much as 80% of total tumor polyamine content.9 We determined the effect of DFMO and MQT 1426, singly and in combination, on polyamine levels in SCCs (Table III). DFMO treatment caused a slow decline in putrescine levels, which did not become significant until the 8-day time point. Combination treatment, however, caused a significant decrease in putrescine content as early as 24 hr after treatment, which persisted for 8 days. Consistent with prior results in SCCs treated with 1% DFMO, neither DFMO treatment alone nor the combination of DFMO plus MQT 1426 had a significant effect on tumor spermidine (Table III) or spermine levels (data not shown).

Table III. Effect of Various Treatments on SCC Polyamine Levels
Treatment GroupTime (d)PuPolyamine nmol/mg DNAp-value1Spmp-Value
  • Tumor-bearing mice were randomized to one of 4 groups based on tumor size distribution. There were 4–9 SCCs per time point. Pu = putrescine, Spd = Spermidine.

  • 1

    vs. control value (2-sided t test).

Control 331 ± 120 160 ± 43 26.1 ± 6.1 
MQT 14261254 ± 1330.45116 ± 70.0520.3 ± 6.10.33
2530 ± 2700.18168 ± 60.7626.4 ± 5.30.98
4509 ± 3030.31262 ± 2070.2127.5 ± 10.50.69
8468 ± 2570.32194 ± 430.2128.7 ± 6.30.63
DFMO1244 ± 1350.58123 ± 380.4418.7 ± 7.10.39
2191 ± 1640.20180 ± 70.6331.5 ± 14.50.55
4408 ± 2920.60174 ± 500.5632.4 ± 18.10.38
879 ± 330.003158 ± 450.9522.9 ± 1.10.48
DFMO + MQT 14261123 ± 600.007130 ± 300.2320.2 ± 8.70.36
2170 ± 1060.01170 ± 630.5129.3 ± 11.50.69
479 ± 630.003241 ± 1390.2227.2 ± 11.90.89
875 ± 670.003164 ± 1210.9430.6 ± 16.30.79

Overall, MQT 1426 treatment alone had no significant effect on putrescine levels over the 8-day treatment period. However, closer inspection of the data indicated a relationship between tumor size, responsiveness to MQT 1426 and putrescine levels (Table IV). As previously mentioned, there is a large variability in SCC growth rates, even for SCCs on the same mouse. Thus, when SCCs in the control groups were stratified based on volume at start of treatment (<500 mm3, 1,000–2,000 mm3, >2,000 mm3) there was a clear relationship between size at beginning of treatment and tumor growth rate: smaller tumors grew faster (258 mm3/week over 4 weeks) and the largest tumors grew slowest (25 mm3/week). Medium size tumors had an intermediate growth rate (94 mm3/week). Similarly, when the combined polyamine data from the 4- and 8-day time points (the earliest times at which a tumor growth response could reliably be detected) were analyzed, there was a clear difference between large vs. small tumors. As shown in Table IV, there was a significant difference in mean putrescine levels between the 5 largest and the 5 smallest SCCs (1,261 ± 601 nmol/mg DNA vs. 277 ± 101; p = 0.03, 2-sided t test). Since smaller tumors are growing at a faster rate, they are likely more dependent on imported (vs. intracellularly synthesized) polyamines than larger, slower growing tumors, and thus are more sensitive to growth inhibition by MQT 1426. In terms of response to MQT 1426 treatment, the 5 largest SCCs increased in volume throughout the treatment period, while 3 of the 5 smallest tumors decreased in volume. A similar analysis of DFMO-treated SCCs did not reveal a tumor volume-related differential sensitivity to growth inhibition (data not shown).

Table IV. Role of Tumor Volume and Putrescine Content in Response to MQT 1426
Volume rankTime (d)Vol (mm3)Δ Vol (mm3)PuTotal PA
  1. A total of 11 SCCs treated for either 4 or 8 days with MQT 1426 were ranked by volume (highest to lowest) at the time of tumor harvest. Changes in volume were calculated from initial pretreatment tumor volumes. Mean putrescine levels (see text) were calculated in the 5 largest SCCs (volume, rank 1–5) and the 5 smallest SCCs (volume rank, 7–11). Pu = putrescine; PA = polyamine.


Effect of DFMO with or without MQT 1426 on tumor proliferation and apoptotic index

In an effort to understand the mechanism by which the combination of DFMO and MQT 1426 caused tumor regression, we measured the proliferation index and fraction of cells undergoing apoptosis in SCCs from each of the treatment groups. Treatment of SCCs with 0.5% DFMO had a rapid effect on tumor proliferation (Fig. 5). At 2 and 4 days after treatment, SCC cell proliferation was significantly inhibited by DFMO (p < 0.01). MQT 1426 treatment had no effect on the proliferation index, nor did it enhance the antiproliferative effect of DFMO. Thus in terms of its effect on tumor-cell proliferation, the effect of the DFMO/MQT 1426 combination was no more effective than that of DFMO alone.

Figure 5.

Effect of treatments on SCC proliferative index. Pieces of the same SCCs used for ODC and polyamine analyses were fixed and processed for measurements of proliferative index as described in Material and Methods. Results shown are the mean ± SE from >1,000 cells counted from 3–5 different tumors. * p value < 0.05 (vs. control Students t test).

With regard to apoptosis, compared to control SCCs, tumors in either the DFMO alone or MQT 1426 alone treatment groups did not show an increase in the number of cells undergoing apoptosis (Fig. 6). However, in the combination therapy group, a significant increase in apoptotic cells (p < 0.01) was observed at 2 and 4 days after treatment. The percentage of TUNEL-positive cells was elevated (nonsignificantly) as early as 24 hr after the combination therapy, reached a maximum at 48 hr (∼300%), followed by a recovery to control level at 8 days. In contrast to its effect on tumor-cell proliferation, the effect of combination therapy on apoptosis did differ from the effect of DFMO alone. The significance of this transient increase in apoptosis following combination therapy to eventual tumor outcome is not at all clear, however.

Figure 6.

Effect of treatments on SCC apoptotic index. Pieces of the same SCCs used for ODC and polyamine analyses were fixed and processed for measurements of apoptotic index as described in Material and Methods. Results shown are the mean ± SE from >1,000 cells counted from 3–5 different tumors. *p value < 0.05 (vs. control Students t test).


There is a critical need for new approaches to therapy of SCCs. For instance, recurrent or inoperable head and neck SCCs are difficult to manage successfully. The idea of targeting the polyamine metabolic pathway as a therapeutic approach is not new. Indeed, the development of DFMO in the late 1970s represented one of the first rationally designed cancer chemotherapy drugs.16 However, results of early6 and more recent17 trials involving DFMO at high doses in patients with advanced disease were generally disappointing. However, in highly malignant brain tumors, combination therapy of DFMO with standard postradiation chemotherapy has significantly increased patient survival compared to standard therapy alone.18 Although there may be multiple reasons why DFMO is ineffective in most treatment settings, failure to achieve a sufficient degree of polyamine depletion in tumor cells is one of the more obvious explanations. One mechanism that would operate to limit polyamine depletion after DFMO treatment is polyamine uptake from the tumor microenvironment. We reasoned that combination therapy with both DFMO and a PTI might alter the kinetics or extent of polyamine depletion, leading to an enhanced therapeutic effect. We report here the first in vivo study investigating combination chemotherapy with DFMO and a PTI. Our initial results of this new approach to polyamine-based therapy in an autochthonous model of murine SCC are encouraging: combination therapy with DFMO and the PTI MQT 1426 was more effective than with either agent alone in producing complete tumor responses and even apparent cures. In 1 trial, 72% of SCCs treated with the drug combination exhibited complete responses compared to 21% of SCCs treated with DFMO alone. In another trial, an apparent cure rate of 50% was achieved with combination therapy compared to a 7.7% apparent cure rate with DFMO alone. To our knowledge, this is the first demonstration of substantial efficacy of polyamine-based therapy involving a PTI in an in vivo tumor model.

As stated above, the rationale for clinical development of polyamine-based therapy of SCC involving polyamine transport inhibition was to augment the efficacy of ODC inhibitors such as DFMO. Nevertheless, PTIs may be effective as single agents in certain settings. Indeed, in the 2 trials we conducted, 32% of SCCs responded to MQT 1426 with growth inhibition (combined data from Tables I and II). Interestingly, SCCs that responded to MQT 1426 were smaller, faster growing, and had lower putrescine content than did the tumors that did not respond (Table IV), suggesting that such tumors have a polyamine requirement for growth that cannot be filled by intracellular synthesis alone. The only other source of polyamines is the tumor microenvironment, from which they can be imported via the polyamine transporter.

Results of polyamine measurements indicated a significantly faster depletion of tumor putrescine levels after combination therapy vs. DFMO alone, although the extent of putrescine depletion after an 8-day period was similar in both treatment groups (∼80% depletion). In contrast, combination therapy had no significant effect on tumor spermidine or spermine levels, suggesting these polyamines are not growth regulatory in SCCs. We have previously speculated that putrescine rather than spermidine or spermine (or total polyamine content) is the polyamine most likely responsible for changes in tumor growth following DFMO treatment.19 Our results based on measurements of total SCC levels of polyamines are subject to uncertainties about the actual “free” intracellular levels of these cations and whether drug treatment alters either the free or total concentration in intracellular compartments such as the nucleus. It is also not known if MQT 1426 accumulates intracellularly to any extent in tumors and whether such accumulation is relevant to its antitumor effect.

We have recently reported that monotherapy with a high concentration of DFMO (1%) in the same SCC model used in this study resulted in a reduced tumor proliferative index as well as a decreased rate of apoptosis.9 Thus we concluded that an increase in apoptosis was not required for an antitumor effect of high-dose DFMO. In the present study, the response of SCCs to combination therapy with a lower dose of DFMO (0.5%) plus MQT 1426 was qualitatively different. Tumor-cell proliferation was inhibited as expected, but there was no difference in the proliferative response of tumors treated with combination therapy vs. tumors treated with DFMO alone. However, there was a difference in the apoptotic index: combination therapy with DFMO and MQT 1426 caused a significant increase in apoptosis (3-fold at 2 days after treatment), while there was no significant change in the tumor apoptotic index after either DFMO alone or MQT 1426 alone. Thus, one added clinical benefit of MQT 1426 with DFMO appears to be an acute induction of apoptosis in tumor cells, albeit a transient response (Fig. 6) and likely not required for a complete tumor response. Since MQT 1426 has no intrinsic apoptosis-inducing properties of its own, it is most likely the perturbation of tumor polyamine pools that occurs when combination therapy is used that is responsible for the transient induction of apoptosis. Several studies in other models have associated polyamine depletion with an induction of apoptosis.20, 21, 22

The rapidity of the response of SCCs to combination therapy is remarkable. As early as 8 days after treatment, much of the apparent tumor mass consists of stromal elements, including fibroblasts and infiltrating leukocytes (data not shown). The increase in apoptotic rate after combination treatment, while significantly elevated at 2 and 4 days, clearly cannot account for the massive loss of SCC cells that occurs over this 8-day period. In a recent study involving treatment of SCCs in this same mouse model with 1–2% DFMO (no MQT 1426), complete tumor responses were observed in the absence of an increased apoptotic rate.9 We reasoned that the antiproliferative effect of DFMO, combined with an unchanged or even increased rate of cell loss due to terminal cell differentiation, was responsible for the therapeutic response. It is likely that a similar mechanism involving reduced proliferation and cell loss via terminal cell differentiation is operative in the response to combination therapy.

The induction of a large fraction of apparent cures with combination therapy contrasts with the usual cytostatic effect of monotherapy with DFMO in other mouse models. Most of the earlier work in mouse models employed human cancer cell lines grown as xenografts in athymic mice,23, 24 while in the K6/ODC (FVB) model, cutaneous SCCs develop at the site of topical application of a genotoxic carcinogen. There is much concern about the predictive value of xenograft models in evaluating new drug candidates. Autochthonous models such as K6/ODC (FVB) may be more reliable models to test new therapies, especially if the drug or drug combination is directed at a pathway known to be deregulated in the tumor being studied (i.e. polyamine metabolism in SCC). Based on the high percentage of complete responses and apparent cures achieved with combination therapy in this study, it is possible that SCCs are unusually sensitive to this therapeutic modality. Another unanswered question is whether our results are unique to this transgenic mouse model of SCC or could be translated to other models, including human SCCs. There are no published reports of DFMO therapy, either alone or in combination with other agents, against any form of human SCCs (i.e. head and neck, cervix and lung). However, human SCC-derived cell lines have been shown to be susceptible to growth inhibition both in vitro and as xenografts in athymic mice.22 In considering future human trials of combination therapy with DFMO and MQT 1426, there are several factors favoring its further development. First, combination therapy with a PTI appears to enhance the antitumor effect of DFMO alone, somewhat neutralizing the negative bias related to the early7 and more recent17 trials with this agent. Second, the therapy is well tolerated by mice, suggesting that it might be used with standard therapeutic regimens in humans, without significant additional toxicity. Finally, lower doses of DFMO can be used while retaining therapeutic efficacy (a dose of 1% DFMO in the drinking water of mice is ∼10 times the highest dose ever used clinically in humans).7 Upon further refinement in preclinical models, the combination therapy described here should be evaluated against human SCCs, either alone or in combination with standard chemotherapy regimens.


We thank Dr. Alexander Muller for helpful comments on the manuscript. We thank the former Ilex Oncology, Inc., for supplying the DFMO used in our studies.