RASGRP4 is a key factor in the KRAS activation mediated by SOS in tumor Y1 adrenocortical cell lines

Abstract

The Y1 mouse adrenocortical carcinoma cell line presents amplification of the KRas oncogene and high-basal levels of KRAS-GTP mediated by the GEF SOS. In this research, we developed a dynamic model based on ordinary differential equations of the KRAS-GTP activation mediated by SOS in Y1 cells, which showed that SOS alone is not sufficient to reach the high-basal levels of KRAS-GTP experimentally observed for this cell line. Interestingly, a modification in this system, which added another GEF in the model, made the model reach the expected levels of KRAS activation, leading to the hypothesis that there was a missing element in this system. To find this missing element, a PCR panel of RasGEFs was performed, and the GEF Rasgrp4 was found to be highly expressed in parental Y1 cell lines, indicating that this was the missing element in the system. Finally, tumor growth assays in Balb/c-NUDE mice with the Y1 cell versus RASGRP4 CRISPR-depleted Y1 cells showed reduced tumor growth and frequency for the RASGRP4-depleted cells.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-42968-0.

Introduction

The Y1 cell line is a mouse-derived adrenocortical carcinoma, an established model for the study of ACTH receptor-mediated steroidogenesis and signaling^18,20. They present amplification of the KRas oncogene¹⁶, with high basal levels of KRAS-GTP, even in the absence of external stimulation³⁰. In addition, Fibroblast Growth Factor 2 (FGF2), known for its role in neurogenesis, morphogenesis, and wound healing¹, impairs the cell cycle in G2/M, restraining the proliferation of the Y1 cell line via the ERK pathway, leading to cell death^8,30. These characteristics make this cell line relevant to study the MAPKs (RAS-RAF-MEK-ERK) pathway, which plays a key role in many cancers^9,25,27.

The tumorigenicity observed in this cell line is a result of a deregulation of the switch mechanism that is controlled by the small GTPases (GEFs and GAPs), mediated by the GEF SOS^5,23, resulting in the aforementioned high KRAS-GTP basal levels. Small GTPases of the RAS superfamily (also called G-Proteins) are part of the GTP-binding proteins, which function as molecular switches, ranging from the GDP-bound inactive state to and GTP-bound active state. This switching mechanism is tightly regulated by two types of accessory proteins: GTPase-activating proteins (GAPs), which promote the hydrolysis of GTP to GDP, switching RAS to its inactive state, and guanine nucleotide exchange factors (GEFs), which facilitate the release of GDP, allowing GTP to bind and activate RAS and consequently their downstream kinases^4,6.

It is also crucial to note that in the Y1 cell line, the amplification of the Kras gene itself is insufficient to drive tumorigenicity, as well as the KRAS activation mediated by SOS. The combination of these two factors must occur to consequently trigger the ERK signaling cascade that underlies this phenotype. Although there is some level of understanding of the SOS mediation of this mechanism, much of the entangled interaction of the other GEFs is yet to be learned¹⁷. To better understand this complex interaction, we have developed a dynamic model based on ordinary differential equations (ODE) of the KRAS-GTP activation mediated by SOS in Y1 cells, which pointed to the need for an additional unknown factor. After careful investigation using classic molecular biology techniques and validation through mouse tumor assay, we report RASGRP4 as a novel necessary GEF to drive tumorigenicity in the Y1 cells, where its absence prevents the tumor growth in Balb/c-NUDE mice. These findings point to the need for greater attention to the GEFs as a whole, a rationale that can be applied in other clinically relevant cancer lineages.

Results

ODE modeling of the steady-state of KRAS activation in Ycells mediated by SOS displays unexpected instability

To construct the dynamic model of KRAS activation in Y1 cells, we chose to use the species and reactions previously described by Jayaji Das et al., in 2009⁹, which takes into account the allosteric regulation of SOS by RAS-GDP and RAS-GTP, which sets up a positive feedback (Fig. 1).

Fig. 1.

Representation of the ODE modeling of the steady-state of KRAS activation in Y1 cells mediated by SOS. Species and chemical reactions that constitute the RAS molecular switch. Reaction rate constants are indicated next to them. Ras alternates between an active state (RAS-GTP), catalyzed by GEFs such as SOS, and an inactive state (RAS-GDP), catalyzed by GAPs. SOS is allosterically regulated by both RAS-GTP and RAS-GDP, which leads to an elevation of its catalytic activity. As allosteric regulation by RAS-GTP increases the catalytic activity of SOS by 75-fold (compared to a 15-fold increase provided by the same regulation by RAS-GDP), the molecular switch presents positive feedback. A: RAS molecular switch with only SOS as GEF. B: The same switch with the inclusion of an additional GEF, which does not present positive feedback.

This model derived the following system of ODEs:

Equation 1

ODE modeling of the steady-state of KRAS activation in Y1 cells mediated by SOS. The Kinetic parameters of this model can be accessed in Table 1.

Table 1.

Kinetic parameters used in this study for modeling Ras activation dynamics in Y1 cells. Rate constants were assigned standard biochemical units based on reaction order. Bimolecular association rates (e.g., k₁, k₂) are expressed in µM⁻¹·s⁻¹, while unimolecular dissociation rates (d₁, d₂) and catalytic rates (k₍cat₎) are in s⁻¹. Michaelis constants (K_m) are reported in µM, and maximum velocities (e.g., V₅) in µM·s⁻¹. These conventions follow classical mass-action and Michaelis–Menten kinetics, as adopted in previous modeling efforts⁹.

Parameter	Description	Value (Y1 cells)
k_1	SOS binding Ras-GDP (kon)	1.8 × 10−⁴ µM−¹·s−¹
d_1	SOS–Ras-GDP dissociation (koff)	3.0 s−¹
k_2	SOS binding Ras-GTP (kon)	1.7 × 10−⁴ µM−¹·s−¹
d_2	SOS–Ras-GTP dissociation (koff)	0.04 s−¹
k_{3cat}	Catalysis with Ras-GTP (allosteric site)	0.038 s−¹
K_{3 m}	Michaelis constant for k_{3cat}	1640 µM
k_{4cat}	Catalysis with Ras-GDP (allosteric site)	0.003 s−¹
K_{4 m}	Michaelis constant for k_{4cat}	9120 µM
V_5	RasGAP-mediated deactivation	0.1 × [GAP] µM·s−¹ approx.
K_{5 m}	Michaelis constant for GAP	107 µM
k_{6cat}	Additional GEF-mediated exchange	0.01 s−¹
K_{6 m}	Michaelis constant for additional GEF	1836 µM

Where SOS-RAS-GDP and SOS-RAS-GTP denote SOS regulated allosterically by RAS-GDP and RAS-GTP, respectively. The rate constants were defined with experimentally measured values found in the literature⁹. Thus, we explored the newly defined model for different total concentrations of KRAS and SOS, increasing the former to up to 20 times the experimentally measured concentration in HeLa cells⁹, in an attempt to mimic the high basal levels of KRAS-GTP observed in Y1 cells. However, we observed that for different total concentrations of KRAS, it was not possible to find a total SOS value that could explain the constitutively high basal levels of KRAS-GTP observed in Y1 cells. The system either entered a stationary state with negligible levels of KRAS-GTP or all KRAS remained in the KRAS-GTP form, thus yielding results inconsistent with those experimentally found in Y1 cells. Therefore, we speculated that the dynamic model was probably missing critical reactions to explain the functioning of the KRAS molecular switch in Y1 cells. Since our group reported years ago the involvement of GEFs in maintaining high basal levels of KRAS-GTP in Y1 cells¹⁴, we decided to include an additional GEF in the dynamic model, whose presence in this cell line was still unknown; this was done by adding the term:

To the right side of Eq. (1e) of the ODE system, and the same term with a negative sign to the right side of Eq. (1d). With these modifications, it was possible to obtain a stable high basal level of KRAS-GTP. It is worth noting that the switch maintained the ability to alternate between “off” and “on”, although it lost the bistability observed in the first trial. (Fig. 2). The kinetic parameters can be seen in Table 1:

Fig. 2.

Elevated basal levels of RAS-GTP dependent on another GEF besides SOS. For a given total RAS concentration, states were computed for different SOS concentrations, taking the RAS-GTP concentration as the output reading. Data are presented on a log scale; units correspond to the original linear values. A: To calculate these steady states, the ODE system was used, counting only SOS as GEF. B: ODE system including an additional GEF.

PCR analysis of YGEFs sheds light on the unknown factor

The results of the ODE modeling prompted the hypothesis of an absent GEF in the SOS-mediated KRAS activation in Y1 cells. To confirm this, RT-qPCR analyses of the RAS GEFs Sos1, Sos2, Grf1, Grf2, Grp1-4 were performed in the presence or absence of FBS, with and without FGF2 stimulation. This growth factor is responsible for increasing ERK activation via the RAS → RAF → MEK → ERK axis, which culminates in cell death for this cell line in a prolonged stimulation as mentioned previously. In contrast to Y1 cells, clones of the FGF2-dependent Y1 cell line (Y1-FD), obtained through selection of cells that survived sustained FGF2 stimulation, were also analyzed. In these cells, the deleterious effects of FGF2 observed in the parental line are absent, and cell cycle progression becomes FGF2-dependent. We hypothesize that this occurs because the classical ERK activation mechanism was disturbed, allowing a comparative analysis of the GEF panel in these two cell types to reveal which GEFs are most relevant in Y1 cells. In deprived Y1 cells, Fig. 3A shows the expression levels of each GEF relative to the level of Sos1, where Rasgrp4 was the predominant GEF in parental cells, being 9.6 times more expressed than Sos1. Figure 3B shows the levels of each GEF in all lineages and growth conditions evaluated, always relative to the level of the respective Ras GEF in starved Y1 cells. The highlight in these gene expression results is that Rasgrp4, the predominant GEF in Y1 cells, is absent in all Y1-FD lines. Therefore, the result points that the high basal-levels of KRAS-GTP in Y1 cells are also dependent on high Rasgrp4 expression. This result prompted us to conclude that RASGRP4 is the GEF that could explain the high KRAS-GTP-bound levels found in Y1 cells, as predicted by the model.

Fig. 3.

Analysis of expression levels of Ras GEFs mRNAs in parental Y1 cells and in the sublines Y1-FD 1 and 2. RT-qPCR analysis of RasGEF expression. mRNA levels of Ras-GEFs (Sos, Grp and Grf families) were quantified by RT-qPCR in Y1 parental cells and Y1-FD clones (1 and 2) under serum-free and exponential growth conditions, with or without FGF2 treatment (10 ng/mL, 24 h). Cells were serum-starved for 24 h prior to treatments. Relative expression was calculated using the Pfaffl method, normalized to Hprt, and expressed relative to the untreated condition. A: mRNA expression levels of Ras GEFs in Y1 cells, normalized to Hprt and ratioed to Sos1 expression. B: Ras GEFs mRNA levels for all cell lines, normalized to Hprt and ratioed to the levels of their respective Ras GEF in serum-deprived Y1. Legend: Dep = Deprived of FBS. Exp+FGF2 = Cells at exponential growth treated with FGF2. Exp-FGF2 = Cells at exponential growth with no FGF2 treatment.

Once the expression of Rasgrp4 was not detected in both Y1-FD clones, we chose to proceed with only the Y1-FD1 clones for further applications.

RasGrpdepletion reduces cytotoxic effects caused by FGF2 in Y1 cells due to lower K-Ras activation

In sequence, the Y1 cell with the CRISPR-depleted Y1 ΔGRP4,Y1 ΔKRAS and Y1-FD cell lines, regarding its proliferation, clonogenic capacity, and expression of genes from the Ras family (Kras, Hras and Nras).

The growth curves (Fig. 4A) showed that the Y1 parental cell line and ΔKRAS cell lines had a slight growth difference, with both having 26.4 and 31.5 h of doubling time, respectively. The ΔRASGRP4 and Y1-FD 1 cell lines showed a shorter doubling time of 14.9 and 12.1 h, respectively. The Ras activation assay (Fig. 4B) has demonstrated differential Pan-Ras-GTP-bound levels among the cells used in this study. Regarding Y1 cells (avg. ABS = 0.619), both ∆KRAS and ∆GRP4 cells exhibited approximately half the activation levels (avg. ABS = 0.2142 and 0.2792, respectively). In contrast, Y1-FD 1 cell displayed negligible activation (avg. ABS = 0.0354). Statistically, Tukey’s test indicated that Y1 cells were significantly different from all other groups (p = 2.5 × 10⁻⁷). Furthermore, ∆GRP4 and ∆KRAS were significantly different from Y1-FD 1 clone cells (p = 1.82 × 10⁻⁵ and p = 5.75 × 10⁻⁴, respectively), but not significantly different from each other (p = 0.288). This assay corroborates the preposition that the RASGRP4, once when removing the newly discovered GEF, it is possible to drastically change the tumor phenotype observed in the Y1 cell line. Lastly, in the clonogenic assay (Fig. 4C), a differential behavior between the lineages in relation to FGF2 can be observed. The ΔKRAS and Y1-FD1 cell lines exhibited resistance to FGF2’s deleterious effect whereas ΔRASGRP4 cells also showed resistance, however, to a lesser extent. In addition, Y1-FD1 cells failed to progress in the absence of FGF2 stimulation.

Fig. 4.

RasGrp4 depletion reduces cytotoxic effects caused by FGF2 in Y1 cells due to lower Ras activation. A: Cell growth curves. Y1 cells were plated in 12-well plates and, after serum starvation (48 h), cultured in DMEM with 10% FBS in the presence or absence of FGF2 (10 ng/mL). Cell numbers were measured over 8 days using an automated cell counter. Doubling time was calculated from the exponential growth phase (days 2–4) using log-transformed cell counts. The individual experiments can be seen in Supplementary Fig. 1. B: Pan-RAS detection through ELISA, comparing all the cell lines from this study. Cells were not synchronized to capture steady-state conditions. Absorbance was measured at 490 nm. Statistical analysis was done with one-way ANOVA followed by Tukey’s post hoc test. Significance: *** = p < 0.001. C: Clonogenic assay of the response to all lineages in this study versus FGF2. For this assay, 600–1000 cells per well were plated on 6-well plates, allowed to adhere overnight, and then treated with 10 ng/mL FGF2. After that, the culture media were replaced every other day until the endpoint of 14 days. The whole experimental set can be seen in Supplementary Fig. 2. Legend: Y1 = Y1 tumor cell. ∆GRP4 = Y1 CRISPR depleted of RASGRP4 protein cell line.∆KRAS = Y1 CRISPR depleted of the KRAS protein cell line. Y1-FD 1 = Y1 FGF2-dependent cell line clone 1.

The RASGRPdepletion led to reduced tumor frequency

The RASGRP4 depletion showed reduced Ras-GTP activation in the Y1 cell, confirming the initial hypothesis that SOS solely is not enough to sustain the high KRAS GTP-bound levels in Y1 cell. Therefore, as the next step the tumorigenic capacity was verified in a tumor growth assay. The assay has demonstrated that the RASGRP4 depletion has changed the tumorigenic phenotype in Y1 cells. For the Y1 cell-injected group, there was a rapid decline in the survival rate in Kaplan-Meier survival curves (Fig. 5A) with the occurrence period ranging from the 13th to and 17th day. Regarding the tumor average (Fig. 5B), the Y1 group shows the most aggressive tumor progression, reaching tumors with approximately 1200 mm³ by the end of the observation period (60 days). In the individual profile (Fig. 5C), 5/6 mice reached tumors higher than 1000 mm³, with the top value of 1666 mm³ and the lowest value of 700 mm³. This profile is expected for this assay, taking into account the high-basal levels of non-mutated Kras in Y1 Cells^21,24,36, acting as a control for this experiment.

Fig. 5.

The RASGRP4 depletion led to lower tumor frequency. Six Balb/c-NUDE mice (5–8 weeks old) per group were subcutaneously inoculated with 5 × 10⁴ cells (Y1, Y1 ΔGRP4, Y1 ΔK-RAS, or Y1-FD 1) in the right flank. Tumor growth and animal weight were monitored every two days. In mice with visible tumors, size was measured every other day. Tumor volume was calculated as V = (L × W × W)/2. The endpoint was defined by 1000 mm³ tumor size in controls or 60 days of observation. A: Kaplan-Meier survival curves for all lineages in this study, the presence of a palpable tumor was considered as an event. N = 6. B: Averaged tumor growth. N = 6. C: Individual tumor growth profile of mice during the observation period. D: Representative images of tumors in mice. Images were acquired on the day of euthanasia, immediately prior to sacrifice, upon reaching the predefined experimental endpoint. The mouse weight during the observation period can be seen in Supplementary Fig. 3. Legend: Y1 = Y1 tumor cell injected mouse.∆GRP4 = Mouse injected with Y1 CRISPR depleted of RASGRP4 protein cell line.∆KRAS = Mouse injected with Y1 CRISPR depleted of the KRAS protein cell line. Y1-FD 1 = Mouse injected with Y1 FGF2-dependent cell line clone 1.

For the ΔKRAS group, the survival rate in Kaplan-Meier survival curves (Fig. 5A) shows slower tumor emergence, with the occurrence period ranging from the 17th to and 42nd day, with a protection rate of 1/6. When it comes to the tumor average (Fig. 5B), the ΔKRAS clone has displayed a mild tumor progression, with approximately 500 mm³ in average by the end of the observation period. In the individual profile (Fig. 5C), only 1/6 mice reached a tumor size larger than 1000 mm³, with the top value of 1183 mm³ and the lowest value of 288 mm³. Even though KRAS is established as the dominant isoform that sustains the tumor phenotype in Y1 cells, it was raised already by¹⁴ that other isoforms, such as HRAS can mitigate the KRAS absence, which might explain the residual tumor. Nevertheless, unveiling the adjacent mechanisms of this residual tumorigenesis still needs further clarification. As for the Y1-FD 1 line, no tumor was observed. Lastly, the ΔRASGRP4 group showed a very low decline in the survival rate in Kaplan-Meier survival curves (Fig. 5A), with the occurrence period ranging from the 28th to and 32nd day, with a protection rate of 4/6. Regarding the tumor average (Fig. 5B), the ΔRASGRP4 clone groups show low tumor progression, reaching tumors with approximately 300 mm³ by the end of the observation period. In the individual profile (Fig. 5C), only 1 out of 6 mice reached a tumor size larger than 1000 mm³, with the top value of 1183 mm³ and the lowest value of 256 mm³. (Fig. 6).

Fig. 6.

Summarization of the methods. A: Initial ODE modeling of KRAS activation mediated by SOS. The levels of activation in the steady state of the model did not match the expected levels. B: After adding another GEF, the levels matched, pointing out that SOS was not solely responsible for the activation. C: Further PCR investigation revealed Rasgrp4 to be not detected in Y1-FD Cells, suggesting it to be the missing GEF in the system. D: A CRISPR against RASGRP4 was performed. E: The mouse tumor assay comparing Y1 and Y1 ΔRASGRP4 cells showed that depleted RASGRP4 Y1 cells cannot progress in tumorigenesis, corroborating the hypothesis that SOS cannot solely activate KRAS in Y1 cells.

Discussion

In this article, we report RASGRP4 protein as a necessary factor alongside SOS in KRAS activation in murine Y1 adrenocortical tumor cells. The ODE system reproducing KRAS activation mediated by SOS resulted in a mismatch with the literature (Figs. 1A and 2A). This prompted us to hypothesize that SOS alone could not be responsible for explaining the KRAS-GTP levels, and another GEF would be needed to stabilize the system. Once accomplished, the KRAS-GTP levels matched the expected levels in the model (Figs. 1B and 2B). Aiming for the missing GEF, we verified the RASGEFs profile of mRNA of the Y1 cell and Y1 FGF2 Dependent cell lines through the qPCR technique. The outcome was that the RASGRP4 was the most expressed in the Y1 cell line (Fig. 3A), whereas not detected in the Y1-FD 1 cell line (Fig. 3B), leading to the hypothesis that RASGRP4 is the missing GEF in the KRAS activation model.

To validate this hypothesis, we compared the ΔRASGRP4 cell line versus the Y1 cell line, which has high expression and activation of KRAS, the ΔKRAS cell line, which has the KRAS depleted, and the Y1-FD 1 cell line, which have no RASGRP4 expression and 50% less KRAS copies¹³. The cell growth curve (Fig. 4A) shows that the CRISPR-depleted cell lines and the Y1-FD 1 cell line are viable and capable of growing, and in some cases, as ΔGRP4 and Y1-FD 1 cell line even surpassing the growing rate of the Y1 cell. It is worth mentioning that the Y1-FD 1 cell line displays a more canonical growth curve: After the 4th day (after the identified log phase), the growth rate declines and ends up reaching a plateau. Differently, the other cell lines continue to show growing tendency at the 8th day, proliferating even after reaching confluence. Moreover, the Y1-FD1 cell line is under constant FGF2 stimulation, which we believe contributes to its high growth rate, in contrast to the inhibitory effect observed in the parental Y1 cell line. The Pan-RAS-GTP ELISA assay showed that the ΔRASGRP4 cells have lower basal levels of Pan-RAS activation in their steady state in comparison to the Y1 cells (Fig. 4B). Given KRAS is the dominating isoform in Y1 cell, we infer that the Pan-RAS-GTP ELISA measured levels are from KRAS in the vast majority. However, a minor role from other isoforms cannot be excluded. In general, the aforementioned results corroborate with the aforementioned ODE modeling results. Regarding the response to FGF2 (Fig. 4C), the ΔRASGRP4 cells were resistant against the aforementioned deleterious effect of the FGF2 versus the Y1 cell, displaying similar results as the ΔKRAS and Y1-FD cell lines. Lastly, the tumor growth assays in Balb/c-NUDE mice with the Y1 cell versus ΔRASGRP4 cells showed reduced tumor growth and frequency for the ΔRASGRP4 cells, displaying lower tumor frequency and size even when compared to the ΔKRAS cell line (Fig. 5C). These results confirm the proposed modeling that shows that SOS cannot solely explain the KRAS-GTP levels found in the Y1 cell (Fig. 1B), and RASGRP4 is a newly reported GEF that fills this system, being fundamental for the tumoral phenotype in this cell line, even not part of the main axis of activation described for this pathway.

RASGRP4 is a member of the RasGRP family²⁸. They are composed of a Ras exchange motif, a CDC25 homology domain, and differently unlike SOS, they possess a diacylglycerol-binding C1 domain and calcium-binding EF hands^12,28. In terms of signaling, the recruitment of RASGRP4 takes a different path, activating Ras and MAPK signaling via its C1 domain in response to the diacylglycerol (DAG)³³. Their importance is reinforced in other research, such as the work of Suiré et al., 2012, which shows that RASGRP4 inhibition impairs the RAS activation in neutrophils in response to the GPCR agonist fMLP, and no other GEF, including SOS, can compensate for this absence³⁴. Moreover, the work of³⁸; showed that the blockage of RASGRP4 considerably reduces diffuse large B cell lymphoma growth³⁸.

Conclusions

Taken altogether, our findings show that RASGRP4 is a newly important GEF for the tumorigenicity in Y1 cells, where its absence reflects a reduced tumorigenic phenotype. Due to the conformational differences from the SOS, we speculate that this also points strongly to the influence of the diacylglycerol-regulated cell signaling pathways concomitantly with the MAPK pathway to drive the tumor phenotype. These findings highlight the importance of studying GEFs, which are often overlooked, and suggest that their roles may be more complex than initially perceived. As the next step, we plan to apply this approach in clinically relevant cell models, also mediated by the GEFs and GAPs. We believe that expanding the knowledge of GEFs and GAPs can significantly advance cancer research, particularly in cancers that depend on this activation mechanism.

Methods

Ordinary Differential Equation Modeling of K-RAS Activation Mediated by SOS in Y1 Cells

To construct the dynamic model of KRAS activation in Y1 cells, we adapted the species and reactions described by Das et al.⁹, which incorporate the allosteric regulation of SOS by both RAS-GDP and RAS-GTP, forming a positive feedback loop. We implemented this reaction scheme as a system of ordinary differential equations (ODEs), with rate constants based on experimentally measured values. The model was then used to simulate KRAS-GTP levels under varying total concentrations of KRAS and SOS, including conditions mimicking the elevated KRAS-GTP basal levels observed in Y1 cells. To determine the steady state of the system of ordinary differential equations (ODEs), all time derivatives are set to zero, reducing the model to a system of algebraic equations whose solutions represent the fixed points. The stability of each steady state is then assessed by computing the Jacobian matrix at the fixed point and analyzing its eigenvalues. If all eigenvalues have negative real parts, the steady state is locally stable, as small perturbations will decay over time.

Cell Lines: The Y1 murine adrenocortical carcinoma cell line was obtained from ATCC. The CRISPR-Cas9 cell lineages were derived from the parental Y1 cell line: The Y1 ΔKRAS cell line was obtained in the work of Dias et al.¹¹, while the Y1 ΔRASGRP4 was obtained during the development of this research. As for the FGF2 Dependent Y1 cell lines (Y1-FD), the clones were selected in our lab after continuous Fibroblast Growth Factor 2 stimulation in the culture media. More details in the FGF2 Dependent Y1 Cells Obtaining Section.

Cell Culture: The Y1 parental and CRISPR-depleted cell lines were grown at 37 °C in 5% CO² atmosphere in Dulbecco’s modified Eagle’s Medium (DMEM; #11885084, Gibco, USA) supplemented with 10% Fetal Bovine Serum (FBS; #S0011, Vitrocell, Brazil). For Y1-FD cell lines, 10 ng/ mL FGF2 was added to the media. Before the execution of all experiments, the synchronization in the phase G0/G1 by serum starvation was performed by the removal of the FBS-supplemented media, PBS A (#14190-136, Gibco, USA) wash, and FBS-free DMEM addition for 24–48 h prior to any stimulation depending on the experiment.

RNA Extraction: The total RNA from cells was extracted with the Illustra RNAspin Mini RNA Isolation kit (#25-0500-72; GE Healthcare) following the manufacturer’s instructions. In brief, the samples were treated with DNase I, and immediately after extraction, the RNA fractions were quantified in the spectrophotometer Nanodrop (Nanodrop, Thermo Fisher Scientific; Waltham, Massachusetts, USA) at wavelengths of 280 nm and 260 nm and stored at -80 °C until use. Next, the complementary DNA (cDNA) was synthesized by reverse transcription with the SuperScriptTM III Reverse Transcriptase kit (#18080085; ThermoFisher Scientific, USA) and used as a template for the real-time PCR reaction.

PCR: To experimentally test the aforementioned computational prediction, we investigated the expression levels of members of the known Ras-GEFs families (Sos, Grp, and Grf) by RT-qPCR, Primers designed using the Primer-BLAST tool³⁷ (Table 2) according to cDNA consensus sequences deposited in GenBank², with specificity checked by performing an in-silico PCR using the UCSC tools¹⁹. Prior to the experiment, the cells were serum-starved for 24 h. The experimental conditions were: (1) deprived of FBS for 24 h, (2) exponential growth (max 70% confluency + FBS) - FGF2 for 24 h, and (3) exponential growth + FGF2 10 ng/ mL for 24 h, for both in Y1 parental and Y1-FD clones 1 and 2. The reaction was performed by using the SYBR^® GREEN PCR Master Mix kit (#4309155; ThermoFisher Scientific) with 40 ng of RT RNA and 2.5 µL 300 nM of forward and reverse primer oligonucleotide solutions, following the manufacturer’s instructions. Amplifications were performed in cycles of denaturation reactions (95 °C, 15 s), annealing, and polymerization (60 °C, 60 s) in a thermocycler (StepOne Plus, Applied Biosystems, USA), using the own StepOne software for data collection. For each sample, the efficiency of the primers was determined by the software LinRegPCR tools (LinRegPCR v. 12.17 software)²⁹, which was used for relative quantification of the data obtained by the Pfaffl method²⁶. All the results were quantified relatively, taking the untreated condition as a comparison. Hprt was used as an endogenous normalizing control for the mRNA expression data. Genes were considered not detected when amplification did not cross the fluorescence threshold within 40 cycles (Ct undetermined). Genes with quantifiable but very high Ct values were considered to be expressed at levels close to the detection limit (Supplemental Table 1). Subsequently, for the analysis of RasGEF expression in Y1, a ratio was performed between the GEFs and SOS1 expression. For the comparative panel of GEFs between Y1 RasGEFs and Y1-FD clones, after normalization, a ratio was performed between the exponential and serum-free conditions for the respective GEFs.

Table 2.

Sequences of primers used for gene amplification by PCR.

Gene	Forward Primer	Reverse Primer
HPRT	GTTGGATATGCCCTTGACTATAATGAG	GGCTTTGTATTTGGCTTTTCCAG
SOS1	AGGTTCAGGGGCAAGTTCAC	AACACGTTCCTCCACATCTGA
SOS2	AGTCCATTGCTGACGGCTTT	CTTCACTGCATGCCTTCAACTTT
GRF1	CGAGAAACAGCGTCATAGCA	AGACTTCAAGGGTGGCTGTC
GRF2	ACCAGCAGCCAAAGGTCATA	GCAGGGAGTCGAGGTTCAAT
GRP1	GTGTTCGAGTGCAAGAAGCG	ATCCTTCTTCGGGTGCATGG
GRP2	AAAGGACTTGGGGGTCCGAA	CAGAGTGCTCGTCATGGTCG
GRP3	GCCATCTTGAGGGGTTCAGG	AGAATGCTCCGAATCCGC
GRP4	GAAAGCCCACGTTCTGTCAC	CCACAGTCCCGACAGCGATA

Y1 ΔRASGRP4 and Y1 ΔKRAS cell lines obtained through CRISPR-Cas9: For this research project, we used the same Y1 cell lines depleted of the Kras gene, which were used in the publication^11,11, and Y1 recently depleted from the Rasgrp4 gene in our laboratory, with both obtained with similar methods. Briefly, we developed five sgRNAs for the respective genes using the CRISPR design tool from MIT (https://www.ensembl.org/index.html; https://crispor.gi.ucsc.edu/)⁷, which can be seen in Table 3. The oligos were cloned into a LentiGuide-Puro plasmid (#52963, Addgene, USA), using the methodology described by Sanjana et al.,³¹. For the lentivirus production, the LentiGuide-Puro constructs, psPAX2 (#12260, Addgene, USA) and pCMV-VSV-G (#8454; Addgene, USA) were transfected into HEK293T cells using Lipofectamine 3000 reagent (#L3000015, ThermoFisher Scientific, USA), following the manufacturer’s recommendations. After 48 h of transfection, the viral supernatants were collected and filtered. The Y1 cells were then cotransduced with the constructs LentiCas9-Blast and LentiGuide-Puro 8 µg/ mL hexadimethrin bromide (#sc-134220, Santa Cruz Biotechnology, USA). After 48 h of transduction, Y1 cells were selected with puromycin 3 µg/ mL and blasticidin 7 µg/ mL for 7 days. Finally, these cells had several clones selected and subcultured, where they were finally tested using the western blot technique with antibodies against the proteins in which their genes were knocked out. Lastly, to obtain individual cells for colony formation, we calculated a serial dilution to distribute 50 cells in three technical replicates across 96-well microplates. Over time, the cells were subcultured sequentially from 96-well plates to 24-well, 12-well, and finally 6-well plates. Once colonies formed, they were transferred to culture flasks for expansion or cryopreserved in a freezing solution (40% FBS, 50% DMEM, and 10% DMSO) at -80 °C.

Table 3.

sgRNA sequences designed for CRISPR targeting of the RasGrp4 gene.

Gene	sgRNA Sequence
GRP4	5’ CACCGAATGTGGTCCCCTCGGCGC 3’
GRP4	5’ AAACGCGCCGAGGGGACCACATTC 3’

FGF2 Dependent Y1 Cells Obtaining: The cell lines used in this study were kindly provided by Dr. Tatiana Guimarães de Freitas Matos. Y1 cells were plated at a density of 6.4 × 10² cells/cm² in triplicate. After plating, the cells were serum-starved for 24 h. before receiving treatment with FGF2 at a concentration of 10 ng/mL. The treatment was maintained with the selection medium containing FGF2, which was renewed three times a week. In the first 24 h of exposure to FGF2, intense cell death was observed, which continued in the subsequent days. After 15 to 20 days of treatment, the emergent colonies were isolated using cloning rings, with the cells being trypsinized and transferred to 12-well plates, where each well contained cells from a single colony. The isolated cells were expanded in the presence of FGF2. The clones obtained were called Y1-FD (FGF2 Dependent). The clones obtained were designated Y1-FD (FGF2-dependent). Two subclones were selected for this study: Y1-FD1 and Y1-FD2. Both subclones were included in PCR analyses, while all other assays were conducted using Y1-FD1 only.

Ras-GTP detection with G-LISA Assay: The colorimetric assay was performed in quintuplicate using the Ras G-LISA Activation Assay Kit (#BK131, Cytoskeleton, USA), according to the manufacturer’s recommendations, and read in a spectrophotometer by measuring absorbance at 490 nm. Unlike the other assays, cells were not synchronized in order to capture steady-state signaling. Cells were collected at approximately 80% confluence, and lysis was performed on ice using Lysis Buffer (60 mM NaCl, 20 mM Tris-HCl, pH 7.5) and samples were immediately stored in liquid nitrogen to avoid GTP degradation and stored until further applications. Lysate concentrations were equalized prior to analysis. Statistical differences between groups were assessed using one-way analysis of variance (ANOVA), followed by Tukey’s post hoc test for multiple comparisons. The analysis was performed using Python with the statsmodels (Seabold et al., 2010)³² and scikit-posthocs³⁵ libraries.

Clonogenic Assay: For this assay, 600–1000 cells per well were plated on 6-well plates, allowed to adhere overnight. Prior to the experiment, the cells were serum-starved for 48 h and then treated with 10 ng/mL FGF2. After that, the culture media were replaced every other day until the endpoint of 14 days. Cells were then washed with PBS, fixed, and stained in a fixing/staining solution (0.05% crystal violet, 4% formaldehyde, 10% methanol in PBS) and washed abundantly. Images were acquired using UVITEC Cambridge (UVITEC, UK).

Growth Curves: For this assay, 0.8 × 10⁴ cells were plated on 12-well plates in triplicate in 3 independent experiments, allowed to adhere overnight. Prior to the experiment, the cells were serum-starved for 48 h and then treated with 10 ng/mL FGF2. The cells were then treated with or without FGF2 in DMEM 10% FBS. In the 8 days, cells were harvested, diluted in PBS, and counted in a Z2 Beckman Coulter^® (Beckman, USA) counter. The log phase of cell growth was identified between days 2 and 4. Doubling time was calculated using cell counts from this interval (48 h). Values were log-transformed, and doubling time was determined using the formula: doubling time = (t × log(2)) / (log(Nt) − log(N0)), where t is the time interval (48 h), N0 is the cell number at day 2, and Nt is the cell number at day 4.

Tumor Growth Assay: The animal experiment was approved by the Ethics Committee on Animal Use of the Butantan Institute of São Paulo (CEUAIB), under protocol number CEUA 3,976,280,723. All procedures were conducted in accordance with Brazilian Law No. 11.794/2008, Decree No. 6.899/2009, and the regulations issued by the National Council for the Control of Animal Experimentation (CONCEA). The study is reported in accordance with the ARRIVE guidelines (https://arriveguidelines.org). A total of 24 Balb/c Nude male mice were used in this study. Animals were obtained from the Central Animal Facility of the Faculty of Medicine of the University of São Paulo and maintained under specific pathogen-free conditions, with controlled temperature and humidity, a 12 h light/dark cycle, and ad libitum access to food and water. Animal age at the beginning of the experiments was 5–8 weeks. Subcutaneous xenograft experiments were performed using the murine adrenocortical tumor cell line Y1 and genetically modified derivatives generated by CRISPR/Cas9, including ΔRASGRP4, ΔKRAS and in addition the Y1-FGF2 Dependent (Y1-FD) cell lines. Cells were inoculated subcutaneously into mice’s right dorsal flank at a concentration of 5 × 10⁴ cells per injection, suspended in 50 µL of saline solution. For Kaplan-Meier survival we considered as an event the visualization of the presence of the first palpable tumor. Tumor growth was monitored over time by external caliper measurements, and tumor volume was calculated using the formula (length × width²) / 2. Animals were regularly monitored for general health, body weight, and tumor burden. To evaluate tumorigenic capacity, mice were inoculated with the aforementioned cell lines, and tumor incidence and growth were compared among groups. Animals were euthanized at predefined humane endpoints in accordance with institutional guidelines. Humane endpoints included a tumor volume reaching 1000 mm³, at which point animals were euthanized on the day following measurement, or a body weight loss greater than 20%, whichever occurred first. Euthanasia was performed by intraperitoneal administration of the anesthetic xylazine (30 mg/kg) and ketamine (270 mg/kg), diluted in saline solution. Mouse assay data were processed with Pandas (McKinney and others, 2011)²² 2.2.3 libraries for calculations, KaplanMeierFitter from Lifelines 0.30.0¹⁰.

Data Processing and Visualization: For generating the figures, we used Matplotlib 3.9.0³ and Science Plots 2.0.0¹⁵ for styling.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

FM: Methodology, Investigation, Data Curation, Writing – Original Draft, Visualization. RAMW: Conceptualization, Methodology, Validation, Investigation, Data Curation, Writing – Review & Editing. TEPT: Methodology, Investigation, Data Curation. KAMT: Methodology, Investigation, Data Curation. CSF: Methodology, Investigation, Data Curation. MSR: Conceptualization, Methodology, Software, Formal Analysis, Data Curation, Writing – Review & Editing, Visualization, Supervision. HAA: Conceptualization, Validation, Formal Analysis, Resources, Data Curation, Writing – Review & Editing, Supervision, Project Administration, Funding Acquisition.

Funding

This work received funding from: FAPESP Grants; #13/07467-1; #20/10329-3, #19/21619-5; #20/08555-5; #19/24580-2; #21/04355-4, CAPES Grant #001 and Butantan Institute.

Data availability

The ODE model script can be found in the https://github.com/anthraxodus/Montoni2025_RASGRP4_KeyGEF_Y1 Github repository. The source data for all the figures are included in this paper.

Declarations

Competing interests

The authors declare no competing interests.

Ethics approval

The animal experiment was approved by the Ethics Committee on Animal Use of the Butantan Institute of São Paulo (CEUAIB), under protocol number CEUA 3976280723. All procedures were conducted in accordance with Brazilian Law No. 11.794/2008, Decree No. 6.899/2009, and the regulations issued by the National Council for the Control of Animal Experimentation (CONCEA). The study is reported in accordance with the ARRIVE guidelines (https://arriveguidelines.org).