Cytoplasmic SNHG1 is Enriched in Prostate Cancer Stress Granules

Abstract

SNHG1 is a long non-coding RNA (lncRNA) with a variably reported subcellular localization to either the nucleus or cytoplasm. Its primary activity has been suggested as a competitive endogenous RNA (ceRNA), however, SNHG1 has also been shown to bind directly to proteins, including G3BP1, G3BP2, and DDX3X, all of which are associated with stress granules. Here we analyze the subcellular localization of SNHG1 in prostate cancer cells and determined the association with stress granules. Quantitation of individual SNHG1 transcripts by single molecule (sm)RNA-FISH showed that C4-2B and PC3 cells contained a median of 13 and 7.5 transcripts per cell, respectively. SNHG1 was predominantly nuclear, with most transcripts found in the nucleus of C4-2B and PC3 cells. The number of transcripts in individual cells varied greatly, from 0 to 36, indicating SNHG1 is expressed at low levels. We induced stress granule formation using thapsigargin to determine whether SNHG1 was localized in stress granules. Stress granules were visualized by G3BP1 immunofluorescence concurrently with SNHG1 RNA by smRNA-FISH. We found enrichment of SNHG1 in C4-2B and PC3 stress granules. Thapsigargin treatment also induced a shift in SNHG1 distribution to favor cytoplasmic localization. In C4-2B cells, the change was small, however, in PC3 cells, SNHG1 localization flipped to become predominantly cytoplasmic. These data show that SNHG1 is predominantly a nuclear RNA expressed at low levels, and stress induction induces an association with stress granules and a cell-dependent shift to the cytoplasm.

Introduction

SNHG1 (Small Nucleolar RNA Host Gene 1) is a non-coding RNA with unclear reported function. We have previously shown that knockdown of SNHG1 results in accumulation of prostate cancer cells in the G₀ (quiescent) phase, and high SNHG1 expression correlates with poor outcomes in prostate cancer patients (Zielske et al. 2023). Its predominant subcellular location is unclear. Shen et al. (2017) and Tian et al. (2024) report that 60–80% of SNHG1 is localized in the cytoplasm of renal, cervical, and colorectal cancer cells. Shen et al. show that cellular distribution shifts to 75–85% nuclear upon treatment with doxorubicin. This shift is suggested to be due to nuclear retention through SNHG1 binding to nucleolin. In contrast, Sun et al. (2017) and Li et al. (2020) report that SNHG1 is found mostly in the nucleus (55–75%) in lung, hepatocellular, and colorectal cancer cells. Localization studies were performed by RT-PCR of nuclear and cytoplasmic fractions or traditional low-resolution RNA-FISH.

The most frequently published function of SNHG1 has been as a competitive endogenous RNA (ceRNA). Overall, SNHG1 has been suggested to interact with dozens of miRNAs, which in turn, regulate dozens of genes. Alternatively, SNHG1 can also bind directly to proteins. SNHG1 has been shown to bind matrin-3 in neuroblastoma (Yang et al. 2019). No mechanistic insight was provided. Data mining their mass spectroscopy results showed that SNHG1 also interacts with G3BP1, G3BP2, PABP1, DDX3X, CAPRIN1, and CCAR1, among others. Each of these proteins are involved in stress granules.

Stress granules are non-membrane bound condensates containing RNA and proteins that form in the cytoplasm of cells after induction of oxidative, endoplasmic reticulum (ER), or chemotherapeutic stress (among others). These structures form rapidly, within an hour, after stress induction, and dissipate rapidly (within an hour) when stress is removed (Wheeler et al. 2016). Stress granules are important modulators of prostate cancer phenotype and drug resistance (Xu et al. 2025). In speckle-type POZ protein (SPOP) mutated prostate cancer, stress granule assembly is enhanced in a CAPRIN1-dependent manner, leading to increased resistance to docetaxel (Shi et al. 2019).

These data, as well as SNHG1 interaction with known stress granule proteins as described above, led us to hypothesize that SNHG1 may be a component of stress granules. We also hypothesized that SNHG1 resides in the cytoplasm. The experiments presented here test these hypotheses. Using smRNA-FISH, we observed SNHG1 RNA, although predominantly nuclear, was also significantly distributed in the cytoplasm. When stress was induced, cytoplasmic SNHG1 was enriched in stress granules and, in PC3 cells, SNHG1 switched to predominantly cytoplasmic localization.

Results and Discussion

SNHG1 subcellular distribution

In previous studies, the location of SNHG1 RNA within the cell was determined using cell fractionization and RT-qPCR, or low-resolution RNA-FISH (Li et al. 2020, Shen et al. 2017, Sun et al. 2017, Tian et al. 2024). These studies reported conflicting dominant subcellular locations. In addition, expression levels were relative. In the current study, we approached the question using smRNA-FISH to directly visualize SNHG1 in cells at single copy resolution (Fig. S1). This also allowed us to determine absolute RNA copy numbers, as well as the distribution of copies on a cell-by-cell basis.

We performed smRNA-FISH for SNHG1 in the prostate cancer cell lines C4-2B and PC3 and counted individual transcripts in each cell. We found that C4-2B cells contained 12.8±4.9 transcripts/cell (median 13), while PC3 cells contained 9.5±7.1 transcripts/cell (median 7.5). The range of expression was broad, with C4-2B cells having 2–21 transcripts/cell and PC3 cells having 0–36 transcripts per cell (Fig. 1A). In contrast, ACTB mRNA was expressed at hundreds of copies (or more) per cell (Fig. S2). These data show that SNHG1 is expressed at very low levels in prostate cancer cells, and some fraction of cells in a population may have no expression at all.

Figure 1: Single-cell analysis of SNHG1 RNA.

SNHG1 RNA in C4-2B and PC3 cells were labeled by smRNA-FISH and manually quantified. A) Histogram representation of SNHG1 distribution per cell. B) Nuclear and cytoplasmic localization of SNHG1 was broken down in histogram form. Bars display the fraction of SNHG1 in the nucleus versus cytoplasm (p<0.001, t test). C) Representative cropped and zoomed images of C4-2B and PC3 cells showing SNHG1 RNA in red fluorescence, stress granules (G3BP1) in green, and nuclei in blue (DAPI). Yellow indicates examples of colocalization of SNHG1 and stress granules and is indicated with arrows. Images are one slice from a z-stack. C) Percentage of cytoplasmic SNHG1 (cSNHG1) that were found to be located in stress granules. For C4-2B (n=57), a total of 434 cytoplasmic SNHG1 RNAs were observed, of which 110 were found in stress granules. For PC3 (n=22), a total of 127 cytoplasmic SNHG1 RNAs were observed, of which 20 were found in stress granules. D) Percentage of stress granules with associated SNHG1. A total of 526 stress granules were observed in C4-2B cells (n=57), of which 110 were found to colocalize with SNHG1. In PC3 cells (n=22), we observed 20 stress granules with SNHG1 out of 214 total stress granules. E) The percentage of nuclear and cytoplasmic SNHG1 was quantified with and without thapsigargin treatment. SNHG1 showed a shift toward cytoplasmic localization after treatment, becoming predominantly cytoplasmic in treated PC3 cells (p<0.0001, Fisher’s Exact test).

Quantitation of RNA located in the cytoplasm versus the nucleus showed that in both C4-2B and PC3 cells, SNHG1 was predominantly a nuclear RNA (Fig. 1B). In C4-2B cells, 81.2±15.0% of SNHG1 RNA was in the nucleus p<0.001, t test). In PC3 cells, 64.6±26.1% of SNHG1 RNA was in the nucleus (p<0.001, t test). This agrees with Sun et al. and Li et al. in the study of lung, hepatocellular, and colorectal cancer (Li et al. 2020, Sun et al. 2017).

Thus, SNHG1 RNA is present at low abundance in prostate cancer cells and displays a predominantly nuclear localization. We cannot exclude the possibility that RNA-RNA interactions may interfere with probe hybridization in the smRNA-FISH procedure and result in underreporting RNA abundance.

SNHG1 co-localization with stress granules

Stress granules form from RNA and proteins in response to stress. Upon stress induction, mRNA translation is inhibited, and some mRNAs are sequestered in stress granules. SNHG1 has not previously been reported to be associated with stress granules, however, analysis of proteomic data by Yang et al. (2019) shows SNHG1 interacts with several stress granule associated proteins. We used G3BP1 as a marker for stress granules and performed SNHG1 smRNA-FISH in conjunction with G3BP1 immunofluorescence to determine whether SNHG1 co-localized with stress granules.

Stress granules were induced in C4-2B and PC3 prostate cancer cells with thapsigargin before smRNA-FISH-labeling and G3PB1 immunofluorescent labeling. Thapsigargin alters Ca homeostasis and induces ER stress and stress granule formation. There were 9–10 stress granules per cell (not shown). We detected SNHG1 RNA colocalizing with G3BP1, indicating SNHG1 was a component of stress granules (Fig. 1C; p<10⁻²⁵, binomial test). SNHG1 RNA was not exclusive to stress granules but also appeared in the cytoplasm and nucleus without stress granule association. Likewise, not all stress granules contained SNHG1. Overall, 25.3% and 15.7% of cytoplasmic SNHG1 was found to be stress granule-associated in C4-2B (n=57) and PC3 (n=22) cells, respectively (Fig. 1D). When SNHG1 was present, it was in all but a few instances, a single copy per stress granule. The proportion of stress granules that contained SNHG1 was 20.9% and 9.35% in C4-2B (n=57) and PC3 (n=23) cells, respectively (Fig. 1E). In contrast, for most mRNAs, <1% of a particular mRNA are found in stress granules (Khong et al. 2017). Untreated cells contained no stress granules. Thus, SNHG1 is enriched in stress granules, compared to many other mRNAs.

The cytoplasmic-nuclear distribution of SNHG1 changed following thapsigargin-induced stress (Fig. 1F). SNHG1 tended to localize to the cytoplasm. In C4-2B cells, 18.4% of SNHG1 was cytoplasmic without stress, but 31.0% was found in the cytoplasm after stress induction. Likewise, 30.8% of PC3 cell SNHG1 was found in the cytoplasm without stress, but 67.6% localized to the cytoplasm after thapsigargin-induced stress. In C4-2B, the cytoplasmic shift was modest, but significant (p<0.0001, Fisher’s Exact test). In PC3 cells, SNHG1 distribution flipped from mainly nuclear localization, to mainly cytoplasmic (p<0.0001, Fisher’s Exact test). This shift was not due to changes in expression. PC3 cells treated with thapsigargin had similar SNHG1 levels to untreated cells and C4-2B cells had slightly higher expression (not shown).

The function of SNHG1 with regards to stress granules is unknown. Through association with G3BP1 and other stress granule proteins, it could play a structural or scaffolding role. While we show that a minority of stress granules at the time point chosen have an association with SNHG1, what is not known is whether this is a transient occurrence. It is possible SNHG1 associates with stress granules, exerts a function, and dissociates. The number of stress granules per cell (9–10) is similar to the number of SNHG1 RNAs per cell (10–13). Additional detailed studies are required to elucidate the significance of SNHG1 and stress granules.

In summary, we show that SNHG1 RNA is present at low copies per cell and is distributed mainly in the nucleus. Induction of stress with thapsigargin results in enrichment of SNHG1 in stress granules and a cell-dependent redistribution of SNHG1 to the cytoplasm.

Supplementary Material

Fig S2

Figure S2: Visualization of ACTB RNA by smRNA-FISH. We used smRNA-FISH to show ACTB RNA in PC3 cells. ACTB is indicated in red, nuclei in blue (DAPI). Image was a z-stack slice and shows hundreds of ACTB mRNA that is mostly cytoplasmic.

FigS1

Figure S1: Single-cell analysis of SNHG1 expression. SNHG1 RNA in C4-2B and PC3 cells were labeled by FISH and manually quantified on a cell-by-cell basis. Representative single z-stack images showing SNHG1 RNA (red) and actin immunofluorescence (green). Nuclei were stained with DAPI (blue).