The anti-Rac1-GTP antibody and the detection of active Rac1: a tool with a fundamental flaw

ABSTRACT

Rac1 is a member of the Rho GTPase family and is involved in many cellular processes, particularly the formation of actin-rich membrane protrusions, such as lamellipodia and ruffles. With such a widely studied protein, it is essential that the research community has reliable tools for detecting Rac1 activation both in cellular models and tissues. Using a series of cancer cellular models, we recently demonstrated that a widely used antibody for visualizing active Rac1 (Rac1-GTP) does not recognize Rac1 but instead recognizes vimentin filaments (Baker MJ, J. Biol. Chem. 295:13698–13710, 2020). We believe that this tool has misled the field and impose on the GTPase research community the need to validate published results using this antibody as well as to continue the development of new resources to visualize endogenous active Rac1.

KEYWORDS:

• Anti-Rac1-GTP antibody
• Rac1
• Rho GTPases
• vimentin
• immunofluorescence
• cancer cells

Introduction

The small GTPase Rac1 is a widely studied protein which has been shown to have essential roles in a variety of cellular processes [1,2]. It functions as a binary switch for downstream signalling events, alternating between an active GTP-bound state and an inactive GDP-bound state, in what is commonly defined as the GTPase cycle. The transition into an active state occurs with the exchange of GDP for GTP, facilitated by Guanine nucleotide Exchange Factors (GEFs). Whereas deactivation occurs due to acceleration of its intrinsic GTPase activity by GTPase Activating Proteins (GAPs) [3,4].

The transient nature of Rac1 activation and the non-covalent interaction with guanine nucleotides make the analysis of its activation status remarkably challenging. The native form of a GTPase bound to GTP must be captured or visualized in any assay aiming to detect activation. In comparison, proteins that undergo covalent modifications in their active state, such as phosphorylation, can be easily detected in the denaturing conditions of a Western blot with antibodies specific for that covalent modification. This is not the case for a GTPase, as denaturing the protein results in nucleotide dissociation. Several techniques have been developed for assessing the activation status of Rac1, including the biochemical assessment of Rac1 in cell lysates using the binding domain of p21-activated protein kinase (PAK) (‘PBD pull-down assays’) [5] and the visualization of Rac1 activation in live cells with FRET biosensors [6,7]. Both techniques have advantages and disadvantages but fail to yield information on the localization of endogenous active Rac1 inside cells or in tissue samples.

A commonly used third option is a commercial anti-Rac1-GTP antibody, a tool which was developed to enable the visualization of endogenous active-Rac1 (http://www.neweastbio.com/GProteinmonoclonalAntibodies/381). This antibody was thought to bind to the native structure of active GTP-bound Rac1, but not that of inactive Rac1, and as such has been widely used for immunofluorescence, immunohistochemistry and immunoprecipitation assays.

The anti-Rac1-GTP antibody detects vimentin rather than active Rac1

Our laboratory recently published work conclusively showing that the anti-Rac1-GTP antibody is not a suitable tool for visualizing active Rac1 [8]. Using a model of PC3 prostate cancer cells, we initially found that the anti-Rac1-GTP antibody immunofluorescence staining colocalized with vimentin staining in a filamentous pattern around the nucleus. We had previously shown that these cells have high levels of active Rac1 in biochemical PBD pull-down assays, making this an ideal model for investigating the validity of this antibody [9]. The staining around the nucleus was intriguing, especially as no colocalization with actin filaments was detected, prompting us to investigate the validity of this antibody further (see Figure1). Our first approach was to generate a Rac1 knockout (KO) PC3 cell line using a CRISPR-Cas9 approach. This cell line was confirmed to be Rac1 null via Sanger sequencing and Western blot. Yet despite this, a similar perinuclear fibrous staining pattern that colocalized with vimentin was detected with the anti-Rac1-GTP antibody in the KO cells. Therefore, we concluded that the signal produced by the antibody is not Rac1 dependent. However, it could be argued that the antibody was still recognizing other Rac isoforms in the Rac1 KO PC3 cells such as Rac3, which is known to be expressed in this cell line [10].

We suspected that creating a pan Rac family knockout would result in non-viable cells and therefore decided to take a different approach by stably depleting the expression of vimentin in PC3 cells. This enabled us to deduce whether the signal from the antibody was or not vimentin-dependent. Stable depletion of vimentin from these cells resulted in a concomitant loss of the anti-Rac1-GTP antibody staining with the loss of the vimentin signal. However, biochemical pull-down assays did not show any reduction in active Rac1-GTP upon vimentin knock-down. Therefore, these results together conclusively show that the signal from the anti-Rac1-GTP antibody was dependent on vimentin expression rather than Rac1 expression.

The anti-Rac1-GTP antibody fails to detect receptor-mediated Rac1 activation

After establishing that the anti-Rac1-GTP antibody does not recognize Rac1 in immunofluorescence assays, we decided to demonstrate the non-specific nature of this antibody beyond any reasonable doubt with its use in a stimulation-dependent context. We initially took advantage of the prostate cancer cell-line LNCaP C4-2. These cells exhibit a large increase in Rac1 activation in PBD pull-down assays upon stimulation with epidermal growth factor (EGF) and do not express vimentin, meaning that any non-specific signal from vimentin would not obscure a signal from the antibody recognizing Rac1-GTP. These experiments revealed that no such increase in signal from the antibody could be detected in EGF-treated cells despite a greater than 20-fold increase in Rac1-GTP levels in the PBD pull-down assay. Surprisingly, there was no signal from the antibody at all in these immunofluorescence assays.

As a second approach, we used a Rac1 FRET biosensor expressed in A549 lung cancer cells, which do express vimentin. We have previously shown that A549 cells form peripheral actin-rich ruffles in response to growth factors [11]. We initially confirmed that these structures are Rac1 dependent by creating a Rac1 KO A549 cell line that was unable to form ruffles in response to EGF. We then demonstrated prominent FRET signals in ruffles in response to EGF. Remarkably, there was no detectable staining from the anti-Rac1-GTP antibody co-localizing with the actin-rich ruffles. Instead, the antibody stained vimentin filaments in EGF-treated A549 cells, recapitulating the localization observed in the PC3 prostate cancer model.

A misleading tool that detects mesenchymal cells

The anti-Rac1 antibody has been used for detecting Rac1 activation in multiple publications over more than 10 years. Based on our results and the lack of reliable validations in previous studies, we believe that this antibody could have led to inaccurate conclusions, particularly when used in a disease context. Vimentin is a widely established marker of cancer cells that have transitioned to a ‘mesenchymal’ state, in which cells acquire highly motile and invasive traits. Coincidentally, migration and invasion are Rac1-dependent, and as such mesenchymal cells may be expected to have more Rac1 activation than those in an ‘epithelial’ state, as we previously reported [12]. Therefore, it may be presumed that the higher signal produced by this antibody in mesenchymal cells compared to epithelial cells is from binding to active Rac1 rather than vimentin. We explored this scenario both in prostate cancer and pancreatic cancer cellular models.

Specifically, we showed that three prostate cancer cell lines with high levels of active Rac1 in PBD pull-down assays (DU145, PC3 and PC3-ML) have strong staining from the anti-Rac1-GTP antibody, whereas four cell lines (BPH-1, LNCaP, LNCaP-C4 and LNCaP-C4-2) that had low levels of active Rac1 in the pull-down assay also had no signal from the antibody staining. This correlation could at first be used to support the argument that the antibody was recognizing active Rac1. However, an additional cell line (RWPE1) did not follow this trend, since it had low levels of active Rac1 in the PBD pull-down assay but in contrast has high vimentin levels. When stained with the anti-Rac1-GTP antibody, RWPE1 cells have prominent filamentous staining despite the low endogenous Rac1-GTP levels, again consistent with the dependency of the antibody signal on vimentin expression rather than active Rac1. A similar correlation between active Rac1 levels in pull-down assays and signal from the anti-Rac1-GTP antibody was also observed in pancreatic cancer cell lines. However, any illusion from this correlation was removed after the stable knockdown of vimentin in PANC-1 cells, which resulted in the loss of the anti-Rac1-GTP antibody signal without a concomitant reduction of active Rac1 in the pull-down assay. Altogether, these examples demonstrate that there are scenarios where this antibody can mislead a researcher into trusting the false information it yields.

Final remarks: lessons from antibody validation

The anti-Rac1-GTP antibody has been used by many for immunofluorescence and immunohistochemistry assays, most of which produced similar perinuclear staining patterns to those we observed as vimentin-dependent in our experiments. Therefore, it is not unlikely that those studies were detecting vimentin staining rather than active Rac1 staining. In addition, this antibody has been used in immunoprecipitation assays to precipitate active Rac1 [13–19]. Our initial approach to validate the antibody included attempting to replicate the immunoprecipitation experiments, yet we were unable to reproduce the detection of Rac1 reported by others in immunoprecipitates. We did retrospectively question whether our immunoprecipitation experiments isolated vimentin rather than Rac1; however, we found that vimentin was insoluble in the lysis buffers we used and therefore precipitated during the clarification of lysates before addition of the antibody.

Immunoprecipitation assays have often been used to validate the antibody for subsequent use in immunofluorescence or immunohistochemistry. This is not appropriate as the detection of precipitated Rac1 is performed with Western blotting, thus not revealing the off-target binding partners of the antibody that would have also precipitated. Our work has focused on the application of this antibody for visualizing active Rac1; therefore, more work needs to be done to investigate whether it is suitable for immunoprecipitation approaches. We encourage any users to thoroughly validate this antibody before using it in such methods.

Some users of the anti-Rac1-GTP antibody have also employed it for probing Western blot membranes [20–22]. We have not assessed the validity of this approach, but according to the widely accepted understanding of how a GTPase is activated it would appear to be fundamentally flawed. As mentioned previously, the activation of Rac1 is dependent on a non-covalent interaction with GTP, which binds into the 3D shape of the nucleotide-binding pocket of the GTPase. The denaturing conditions of a Western blot would result in the loss of the tertiary structure of the protein and subsequent dissociation of GTP. Therefore, for the antibody to work for Western blotting it could only work, theoretically, under non-denaturing conditions.

The inability of this antibody to detect Rac1 in immunofluorescence assays creates a major gap in the toolkit available to researchers for studying this GTPase. There are other techniques/tools that provide robust and reliable ways of assessing Rac1 activation status, but none allow the researcher to visualize endogenous active Rac1 in cells or tissue samples. Biochemical PBD pull-downs of active Rac1 are cheap to perform but only yield information on the proportion of Rac1 that is active in a population of cells. Although not highly sensitive, G-LISA assays also yield information from a population of cells and can be applied in a relatively high-throughput manner. FRET biosensor assays can be used to obtain information on the spatiotemporal activation of Rac1 at a cellular level. Limitations include the need for overexpression of the biosensor as well as its use only in cell lines and transgenic organisms but not in fixed samples. Due to the weaknesses of each technique, the best practice approach would be the use of multiple strategies for measuring Rac1 activation. This would help ensure that any artefacts from one of the techniques does not lead to false conclusions. The advantages and disadvantages of the main techniques for measuring Rac1 activation dynamics are summarized in Figure2.

In summary, the GTPase community is left with the pressing need for a tool that can measure endogenous active Rac1, particularly to be applied in human or mouse tissue samples and specimens. The inability to visualize endogenous Rac1-GTP is not new, but our awareness of this shortcoming is. How this can be addressed in the future is not yet apparent, but it has become clear that any new approaches for measuring active Rac1 or for measuring the dynamics of other GTPases, need to be thoroughly validated before use.Figure 1

Figure 1. Staining from the anti-Rac1-GTP antibody colocalizes with vimentin. Representative image of PC3 cells that have been fixed and stained for vimentin (green), DNA (blue), and with the anti-Rac1-GTP antibody (red). Scale bar, 10 μm.

Figure 2

Figure 2. Summary of the main methods available for the measurement of active Rac1.

Acknowledgments

M.G.K. wants to acknowledge funding from grants R01 CA189765, R01 CA196232 and R01 ES026023 from NIH.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by the National Institutes of Health [R01 CA189765]; National Institutes of Health [R01 ES026023]; National Institutes of Health [R01 CA196232].