Distinct requirements for Ras oncogenesis in human versus mouse cells

Abstract

The spectrum of tumors associated with oncogenic Ras in humans often differs from those in mice either treated with carcinogens or engineered to sporadically express oncogenic Ras, suggesting that the mechanism of Ras transformation may be different in humans. Ras stimulates primarily three main classes of effector proteins, Rafs, PI3-kinase, and RalGEFs, with Raf generally being the most potent at transforming murine cells. Using oncogenic Ras mutants that activate single effectors as well as constitutively active effectors, we find that the RalGEF, and not the Raf or PI3-kinase pathway, is sufficient for Ras transformation in human cells. Thus, oncogenic Ras may transform murine and human cells by distinct mechanisms, and the RalGEF pathway--previously deemed to play a secondary role in Ras transformation--could represent a new target for anti-cancer therapy.

Figure 1

Ras12V37G transforms human cells. (A) Expression of Ras mutants was confirmed in NIH 3T3 or HEK-HT cells stably infected with retroviruses encoding the described FLAG-tagged H-Ras cDNAs or with a vector or H-Ras control, by immunoblotting with an anti-FLAG antibody. Actin levels serve as a loading control. (B) Anchorage-independent growth of NIH 3T3 (black bars) or human HEK–HT (white bars) cells expressing the described constructs, calculated from the average number of colonies observed from three plates and expressed as the percent of colonies observed in Ras12V-transformed cells. A total of 50,000 Ras12V-transformed NIH 3T3 or HEK–HT cells yielded 380 ± 50 or 289 ± 47 colonies in soft agar, respectively.

Figure 2

Ras-induced transformation is different in human vs. murine cells. (A) Expression of T-Ag was confirmed in human and murine fibroblasts stably infected with a retrovirus encoding the T-Ag gene by immunoblotting with an anti-T-Ag antibody. Actin serves as a loading control. (B) Expression of Ras mutants was confirmed in T-Ag expressing murine and human fibroblasts (the latter of which also ectopically express hTERT) stably infected with retroviruses encoding the described FLAG-tagged H-Ras cDNAs or with a vector control, by immunoblotting with an anti-FLAG antibody. Actin levels serve as a loading control. (C) Anchorage-independent growth of the aforementioned T-Ag-expressing human (black bars) and murine (white bars) fibroblasts, calculated from the average number of colonies observed from three plates and expressed as the percent of colonies observed in Ras12V-transformed cells. A total of 50,000 Ras12V-transformed human or mouse fibroblasts yielded 320 ± 100 or 481 ± 135 colonies in soft agar, respectively. (D) Anchorage-independent growth of a different independently derived mouse fibroblast cell strain stably infected with constructs encoding T-Ag and the described H-Ras mutants, calculated from the average number of colonies observed from three plates and expressed as the percent of colonies observed in Ras12V-transformed cells. A total of 50,000 Ras12V-transformed mouse fibroblasts yielded 497 ± 64 colonies in soft agar.

Figure 3

Ras12V37G transforms human astrocytes. (A) Expression of Ras mutants was confirmed in hTERT, T-Ag-expressing human astrocytes stably infected with retroviruses encoding the described FLAG-tagged H-Ras cDNAs or with a vector control, by immunoblotting with an anti-FLAG antibody. Actin levels serve as a loading control. (B) Anchorage-independent growth of the aforementioned hTERT, T-Ag human astrocytes expressing the described constructs, calculated from the average number of colonies observed from three plates and expressed as the percent of colonies observed in Ras12V-transformed cells. A total of 50,000 Ras12V-transformed human astrocytes yielded 120 ± 40 colonies in soft agar.

Figure 4

Specific activation of Raf or PI3-kinase is not sufficient to promote anchorage-independent growth of human cells. (A) Expression of ΔRaf1-22W or p110-CAAX was confirmed in the described HEK–HT cells by immunoblotting with an anti-Raf1 antibody or by RT–PCR with primers specific for p110. Actin and GAPDH levels serve as loading controls. These cells also were infected with an empty vector. (B) Specific activation of either the MAP-kinase or PI3-kinase pathway by ΔRaf1-22W or p110-CAAX, respectively, was confirmed in the described HEK–HT by immunoblotting with antibodies specific for the phosphorylated forms of ERK1 and ERK2 or Akt. Total ERK 1/2 and Akt levels serve as a loading control. (C) Anchorage-independent growth of HEK–HT cells expressing ΔRaf1-22W or p110-CAAX, as calculated from the average number of colonies observed from three plates, and expressed as the percent of colonies observed in Ras12V-transformed cells. A total of 50,000 HEK–HT cells expressing H-Ras12V seeded in soft agar yielded 311 ± 46 colonies.

Figure 5

The MAP-kinase and PI3-kinase pathways together are not transforming, but greatly enhance Ras12V37G-mediated transformation. (A) Coexpression of Ras12V effector mutants was confirmed from direct sequencing of H-Ras cDNA to identify the point mutations giving rise to 40C (green), 37G (blue), or 35S (red) effector mutations or the corresponding nucleotide of endogenous H-Ras (black in vector control) in the described double-infected HEK–HT cells. (B) Anchorage-independent growth of HEK–HT cells expressing the described combinations of Ras12V mutants with control vector (black bars) or the described effector mutants, ΔRaf1-22W or p110-CAAX (white bars), as calculated from the average number of colonies observed from three plates, and expressed as the percent of colonies observed in Ras12V+vector-transformed cells. A total of 50,000 Ras12V+vector-transformed HEK–HT yielded 295 ± 83 colonies in soft agar. (C) Anchorage-independent growth of HEK–HT cells expressing Ras12V and control vector, vector alone, or ΔRaf1-22W and p110-CAAX, as calculated from the average number of colonies observed from three plates, and expressed as the percent of colonies observed in Ras12V+vector-transformed cells. A total of 50,000 Ras12V+vector-transformed HEK–HT yielded 463 ± 55 colonies in soft agar.

Figure 6

RalGEFs are essential for Ras transformation of human cells. (A) RalA-28N and Rlf-CAAX expression was confirmed in HEK–HT or Ras12V-transformed HEK–HT cells, respectively, by immunoblotting with either an anti-Ral antibody (top) or an anti-HA antibody specific for the HA epitope cloned into Rlf (bottom). Actin levels serve as loading controls. (B) Phosphorylated Akt (P-AKT) or ERK 1 and 2 (P-ERK 1/2) levels were determined in the described cells by immunoblotting with antibodies specific for the phosphorylated forms of these proteins. The total amount of Akt, ERK 1/2 or actin detected by anti-Akt, ERK, or actin antibodies serve as loading controls. (C) Ral–GTP levels were determined in the described cells by binding cellular Ral–GTP to the recombinant Ral-binding domain of the RalBP protein, followed by immunoblotting with a RalA-specific antibody. The total amount of Ral detected by the same anti-RalA antibody serve as loading controls. (D) Anchorage-independent growth of Ras12V or Ras12V37G-transformed HEK–HT cells in the absence (vector) or presence of RalA28N, or HEK–HT cells in the absence (vector) or presence of Rlf-CAAX, as calculated from the average number of colonies observed from three plates, and expressed as the percent of colonies observed in Ras12V (black bars) or Ras12V37G (white bars) transformed cells. A total of 50,000 Ras12V-transformed HEK–HT cells yielded 295 ± 83 colonies in soft agar.

Figure 7

Ras12V37G is essential for tumor growth. Tumor volume (mm³) of four mice injected with HEK–HT cells expressing H-Ras12V40C and H-Ras12V35S either in the absence (vector: ○□▵⋄) or presence (●▪▴♦) of H-Ras12V37G, plotted against time (days).