Methionine addiction is a fundamental and general hallmark of cancer, known as the Hoffman effect (1-12). Cancer cells show an increased demand for methionine to maintain their proliferation and survival, whereas normal cells can still grow under methionine-restricted conditions (1,3,4,11). The specific vulnerability of cancer cells to methionine restriction is a therapeutic target and methionine depletion strategies - including methionine-restricted media or diet, and treatment with the methionine-cleaving enzyme recombinant methioninase (rMETase)- have been widely studied to target methionine addiction.

In our earlier studies, we demonstrated that HCT-116 human colon-cancer cells are more sensitive to rMETase treatment than Hs27 normal fibroblasts, showing a significant decrease in viability of HCT-116 cells under methionine depletion, compared to the normal fibroblasts (13). We also showed that HCT-116 cells treated with rMETase underwent cell toxicity but were rescued and resumed growth following methionine replenishment (14).

To better model the tumor microenvironment, we developed a co-culture model of HCT-116 cancer cells with Hs27 normal fibroblasts. In this co-culture model, rMETase treatment selectively eliminated the cancer cells while sparing the normal fibroblasts (15).

Based on these findings, the present study investigated whether methionine replenishment after treatment with rMETase, at the HCT-116 IC50 concentration, could rescue cancer cells in the co-culture model of HCT-116 and Hs27 cells. The present study will enable further understanding of the dynamics of methionine addiction in cancer cells and its potential therapeutic implications under more physiologically-relevant conditions.

Cell culture. HCT-116 human colorectal cancer cells and Hs27 normal fibroblasts were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA). Both cell lines were cultured in Dulbecco’s Modified Eagle Medium (DMEM; GIBCO, Grand Island, NY, USA) supplemented with 10% fetal bovine serum (FBS; GIBCO) and 1 IU/ml of penicillin-streptomycin (GIBCO). Cells were cultured at 37˚C in a humidified incubator under 5% CO2 atmosphere.

Co-culture model. To establish the co-culture model, 3.75×104 HCT-116 cells and 3.75×104 Hs27 cells were seeded together per well in 6-well plates containing DMEM, for a total of 7.5×104 cells per well.

Reombinant methioninase (rMETase) production. rMETase, a methionine-degrading enzyme originally derived from Pseudomonas putida, was produced by fermenting genetically engineered Escherichia coli carrying the Pseudomonas putita methioninase gene. rMETase production and purification were carried out at AntiCancer Inc. (San Diego, CA, USA) as previously described using a heat step, polyethylene-glycol precipitation, and DEAE Sepharose ion-exchange chromatography (16,17).

Recombinant methioninase (rMETase) treatment of co-cultures. On day 2 after seeding, wells were designated as either a control group (no treatment) or a treatment group, which received rMETase at the HCT-116 IC50 concentration of 0.46 U/ml, as determined in preliminary experiments (15). The co-cultures were observed and photographed using an IX71 inverted microscope under phase contrast (Olympus, Tokyo, Japan) on days 2, 4, 6, 8, 10, and 12 after rMETase treatment.

Methionine replenishment of rMETase-treated co-cultures. On day 12, when no viable cancer cells could be observed under phase-contrast microscopy in the treatment group, the culture medium in all wells (with or without rMETase) was replaced with standard DMEM containing methionine, to initiate methionine replenishment in the treatment group. Following methionine replenishment, the co-cultures were observed under phase-contrast microscopy and documented on days 3, 6, and 9 after the medium was replaced with fresh DMEM.

By day 2 of the experimental timeline (day 4 after seeding), HCT-116 cancer cells in the co-culture without rMETase treatment (control group) had already proliferated rapidly and become the dominant cell type, occupying a significant portion of the well surface. Subsequently, in the control group, the clusters of HCT-116 cells continued to expand and became increasingly dominant by day 12, while the Hs27 fibroblasts appeared less prominent, forming narrow streams between the expanding HCT-116 clusters as they were progressively displaced. In the rMETase-treated co-culture (treatment group), the Hs27 normal fibroblasts maintained their viability throughout the observation period. In contrast, HCT-116 cancer cells gradually decreased in number, with no viable cells detectable under phase-contrast microscopy by day 12 after rMETase treatment began (Figure 1). This observation was consistent with and fully reproduced our previous findings (15).

On day 12, the existing medium in both groups was replaced with fresh DMEM containing methionine (methionine replenishment). In the previously-treated rMETase group, viable HCT-116 cancer cells reappeared by day 3 after methionine replenishment (day 15 after initial treatment). By day 6, the proportion of rescued HCT-116 cells became comparable to Hs27 cells, and by day 9, HCT-116 cells had proliferated extensively, becoming dominant over the Hs27 fibroblasts. In the control group, HCT-116 cells continued to expand throughout the observation period, and by day 9 after methionine replenishment, Hs27 cells were no longer visible, with HCT-116 cells dominating the co-culture (Figure 2).

The co-culture system modeling competitive interactions between HCT-116 colorectal cancer cells and Hs27 normal fibroblasts highlights the large disparity in methionine requirements between cancer and normal cells. Treatment with rMETase at the HCT-116 IC50 (0.46 U/ml) shifted dominance toward the normal fibroblasts, such that by day 12 of treatment, no viable cancer cells were detectable by phase-contrast microscopy. However, when rMETase was discontinued and methionine was replenished with DMEM on day 12, HCT-116 cells reappeared by day 15 (3 days post-replenishment) and regained dominance by day 21 (9 days post-replenishment). These cellular kinetics indicate that methionine restriction can reverse the growth advantage of methionine-addicted cancer cells, whereas restoration of methionine permits rapid rescue and regrowth of the cancer population.

Our prior monoculture and co-culture studies established two consistent features: 1) HCT-116 cancer cells are much more sensitive to methionine depletion than Hs27 normal fibroblasts (13,15); and 2) HCT-116 cells exposed to their IC50 dose of rMETase can be rescued by subsequent methionine repletion (14). The present study extends these observations by demonstrating that the same rescue phenomenon occurs in co-culture, where cell-cell competition and contact influence growth dynamics. Thus, the “Hoffman effect” (methionine addiction) is not only evident in isolated cancer cells but also determines cell-population outcomes during rMETase treatment in competition with normal fibroblasts.

In the present study, under IC50-level rMETase, HCT-116 became undetectable and subsequently recovered after methionine repletion, indicating complete functional rescue in the competitive context. The present findings demonstrate both the strong dependence of cancer cells on methionine and their selective depletion by rMETase as well as the rapid rescue and re-expansion that occur upon restoration of methionine. The present findings are consistent with rMETase-induced interruption of methionine supply causing a reversible cell-cycle arrest that can return to active cell cycling upon methionine replenishment (18,19). In co-culture, rMETase selectively inhibits cancer-cell proliferation and survival while sparing normal fibroblasts (15).

Methionine addiction reflects increased transmethylation demand and dependence on sustained methyl-group flux in cancer cells (5,9,11,12,22). Therefore, when methionine is depleted by rMETase, the insufficient methyl-donor supply is expected to induce a reversible cell-cycle arrest, and this is supported by the observed “rescue phenomenon” following restoration of methionine as seen in the present study. This is consistent with studies indicating that depletion of extracellular methionine impairs methyl-dependent epigenetic regulation in methionine-addicted cancer cells, causing them to stop proliferating, whereas normal cells are mostly spared (9-11,20-25).

However, our prior studies have shown that prolonged, high-concentration rMETase can drive HCT-116 cells beyond a recoverable state, so that restoration of methionine can no longer rescue cancer cells (methionine-depletion catastrophe) (14). High concentrations of rMETase can also compromise the growth and division of normal fibroblasts (13). The present findings identify a therapeutic window in which rMETase influences cell-population outcomes by selectively inhibiting methionine-addicted cells while preserving normal cells co-cultured with the cancer cells. Accordingly, we used the HCT-116 IC50 dose (0.46 U/ml) to inhibit the growth of the cancer population without impairing normal-cell growth. Under these conditions, we evaluated how methionine depletion and replenishment affected co-culture dynamics.

The present results highlight two key translational implications. First, maintaining methionine restriction is crucial: interruption, even once cancer cells are undetectable, can rapidly rescue the rare cancer cells in the co-culture and restore their competitive dominance within days. Second, methionine restriction via diet and/or rMETase may serve as a platform that targets cancer-cell-specific vulnerabilities revealed under methionine restricted conditions (1).

Methionine restriction induces the cancer cells to be more sensitive to chemotherapy including: 1) inhibitors of folate and purine metabolism (5-fluorouracil with or without folinic acid, methotrexate); 2) agents targeting methylation-dependent pathways (azacitidine, decitabine); 3) cytotoxic cell-cycle agents (doxorubicin, cisplatinum, gemcitabine, eribulin, paclitaxel); and 4) mTOR-pathway targeted agents (rapamycin) (21).

Treatment of HCT-116/Hs27 co-cultures with rMETase at the HCT-116 IC50 (0.46 U/ml) eradicated cancer cells to undetectable levels by day 12. However, methionine replenishment led to cancer-cell reappearance within three days and renewed dominance by nine days. These findings indicate that continuous methionine restriction is required to maintain normal-cell dominance and to prevent cancer-cell rescue and regrowth. Clinically, the growth kinetics of the methionine-rescued, rMETase-treated cancer cells suggest that methionine-restriction in cancer patients should be maintained without interruption to prevent cancer regrowth, since restoration of methionine availability may result in cancer progression. On the basis of the present findings, future studies will evaluate strategies with methionine restriction to prevent cancer-cell regrowth under clinically-relevant conditions.

rMETase is effective because it targets the fundamental hallmark of cancer, methionine addiction (1-19,22-42). rMETase is showing clinical promise (43-47).

The Hoffman effect of methionine addiction is stronger than the Warburg effect of glucose addiction, as shown by comparison of methionine-based and glucose-based PET imaging, respectively (47,48).

All Authors have no conflicts of interest or financial ties to disclose related to this study.

BMK and RMH conceived the project and designed the study. QH and SL produced rMETase. BMK conducted all experiments and prepared the initial draft. RMH made substantive revisions. JSK, KM, YA, YM, and MB critically reviewed the final manuscript.

This paper is dedicated to the memory of A.R. Moossa, MD; Professor Philip Miles; Sun Lee, MD; Richard W. Erbe, MD; Professor Milton Plesur; Professor Gordon H. Sato; John Littlefield, MD; Professor Li Jiaxi; Masaki Kitajima, MD; Shigeo Yagi, PhD; Jack Geller, MD; Joseph R. Bertino, MD; J.A.R. Mead, PhD; Eugene P. Frenkel, MD; John Mendelsohn, MD; Professor I.J. Fidler; Professor Lev Bergelson; Professor Sheldon Penman; Professor John R. Raper; Professor J.D. Watson; and Joseph Leighton, MD.

The present study was funded by the Robert M. Hoffman Foundation for Cancer Research.

No artificial intelligence (AI) tools, including large language models or machine learning software, were used in the preparation, analysis, or presentation of this manuscript.