NKX3.1 Localization to Mitochondria Suppresses Prostate Cancer Initiation

Abstract

Mitochondria provide the first line of defense against the tumor-promoting effects of oxidative stress. Here we show that the prostate-specific homeoprotein NKX3.1 suppresses prostate cancer initiation by protecting mitochondria from oxidative stress. Integrating analyses of genetically engineered mouse models, human prostate cancer cells, and human prostate cancer organotypic cultures, we find that, in response to oxidative stress, NKX3.1 is imported to mitochondria via the chaperone protein HSPA9, where it regulates transcription of mitochondrial-encoded electron transport chain (ETC) genes, thereby restoring oxidative phosphorylation and preventing cancer initiation. Germline polymorphisms of NKX3.1 associated with increased cancer risk fail to protect from oxidative stress or suppress tumorigenicity. Low expression levels of NKX3.1 combined with low expression of mitochondrial ETC genes are associated with adverse clinical outcome, whereas high levels of mitochondrial NKX3.1 protein are associated with favorable outcome. This work reveals an extranuclear role for NKX3.1 in suppression of prostate cancer by protecting mitochondrial function. SIGNIFICANCE: Our findings uncover a nonnuclear function for NKX3.1 that is a key mechanism for suppression of prostate cancer. Analyses of the expression levels and subcellular localization of NKX3.1 in patients at risk of cancer progression may improve risk assessment in a precision prevention paradigm, particularly for men undergoing active surveillance.See related commentary by Finch and Baena, p. 2132.This article is highlighted in the In This Issue feature, p. 2113.

Figure 1.. NKX3.1 protects against mitochondrial oxidative stress

(A-H) Analyses of Nkx3.1 mutant mice. (A) Strategy. Nkx3.1 mutant (Nkx3.1^−/−), but not wild-type (Nkx3.1^+/+) mice develop PIN by 12 months of age. Cohorts of mice received paraquat (10mg/kg/day in drinking water) or vehicle (water alone) starting at 3 months of age and were sacrificed at 4 months to measure ROS levels or 12 months for phenotypic analysis. (B) Quantification of DHE fluorescence from the anterior prostate of Nkx3.1^+/+ or Nkx3.1^−/− mice treated with paraquat (Par) or vehicle (Veh) for 1 month (n=9-14 mice/group). (C) Histopathology showing hematoxylin and eosin (H&E) staining and immunostaining for Nkx3.1 and γH2AX of anterior prostate from mice treated with paraquat or vehicle for 9 months and analyzed at 12 months of age. Scale bars represent 50μm (low power) or 20μm (high power). Shown are representative images from analyses of 18-28 mice per group. (D) Quantification of the PIN phenotype showing the relative percentage of low-grade PIN (PIN1/2) and high-grade PIN (PIN3/4) in anterior prostate. Data show the summary from analysis of 18-28 mice/group and are expressed as the mean percentage ± SD of the control; P values were calculated using chi-squared (χ²) test. ns, not significant. (E-G) Electron microscopy of mitochondria from anterior prostate of Nkx3.1^+/+ and Nkx3.1^−/− mice treated with paraquat (10 mg/kg/day) or vehicle for 9 months and analyzed at 12 months of age. (E) Representative micrographs; scale bars represent 100nm. (F) Quantification of relative mitochondrial area. (G) Quantification of relative number of mitochondria. Panels F and G represent analyses of 60+ images/mouse and 4 mice/group and expressed as the mean percentage ± SD. (H) Quantification of mitochondrial membrane potential as detected by TMRE fluorescence from mice treated with paraquat (10 mg/kg/day) or vehicle for 9 months and analyzed at 12 months of age. (I-W) Analyses of human prostate cells in culture. (I) Strategy. LNCaP, BPH1, or RWPE1 were treated with paraquat or MitoParaquat to induce general cellular or mitochondrial-specific reactive oxygen species (ROS), respectively, alone or together the mitochondrial-specific anti-oxidant, MitoQ. Production of general cellular ROS or mitochondrial ROS was measured using the dyes indicated. (J-W) Oxidative stress was induced by treating cells with paraquat (100μM for 24 hours), hypoxia (1% pO₂ for 24 hours) or hydrogen peroxide (200μM for 6 hours). (J, K) Western blot analyses of total protein lysates in LNCaP cells expressing 2 independent shRNA for NKX3.1 (shNKX3.1#1, shNKX3.1#2) or the control (shControl) (J) and RWPE1 cells expressing NKX3.1 or the control vector (K). Note that subsequent data show shNKX3.1#1. (L, M) Western blot analyses of total protein lysates in LNCaP cells expressing shRNA for NKX3.1 (shNKX3.1) or the control (shControl) (L) and NKX3.1-expressing (or control) RWPE1 cells (M) treated with paraquat (100μM for 24 hours). γH2AX is a marker of DNA damage. (N-R) Analysis of LNCaP cells. (N) Representative histograms from ROS analyses. (O and P) Quantification of mean ROS intensity as detected by DHE (O) or MitoSOX (P) production. (Q) Quantification of mean mitochondrial mass/density as detected by MitoTracker. (R) Quantification of mitochondrial membrane potential as detected by TMRE fluorescence. (S) Representative histograms from ROS analyses. (T, U) Quantification of mean ROS intensity as detected by DHE (T) or MitoPY1 (U) production. (V) Quantification of mean mitochondrial mass/density as detected by MitoView. (W) Quantification of mitochondrial membrane potential as detected by TMRE fluorescence. Panels N-W show representative data from 3 independent experiments; each experiment was done in triplicate or quadruplicate (9-12 independent samples/group). Unless otherwise indicated, shown is the mean ± SD; P values were calculated using two-sample unpaired Welch t-test. ns, not significant. See also Supplementary Figures S1-S5 and Supplementary Tables S1- S4.

Figure 2.. NKX3.1 localizes to mitochondria in response oxidative stress

(A-E) Localization of NKX3.1 to mitochondria. Panels A and B, show analysis of anterior prostate from Nkx3.1^+/+ mice treated with paraquat (10mg/kg/day in drinking water) or vehicle (water alone) for 9 months. Panels C and D show analysis of LNCaP cells treated with paraquat (100μM) or vehicle for 24 hours. (A,C) Confocal images. Samples were co-stained with an antibody specific for mouse or human NKX3.1 (red) and anti-ATPB to visualize mitochondria (green); nuclei were visualized by labeling with DAPI (blue). Scale bars in Panel A represent 50μm (left and center) or 200μm (right) and in Panel C 100μm. (B,D) Western blot analyses following biochemical isolation of nuclear or mitochondrial fractions, as detected using histone H3 (H3) and ATPB, respectively. (E) Representative immuno-EM photomicrographs from LNCaP cells to detect endogenous NKX3.1 or NKX3.1-expressing RWPE1 cells to detect the exogenous protein. Cells were treated with paraquat (100μM) or vehicle (media alone) for 24 hours followed by detection of NKX3.1 (black dots) in mitochondria using an anti-NKX3.1 antibody. Scale bars represent 100nm (left and center) or 400nm (right). (F-P) Analyses of NKX3.1 polymorphisms associated with cancer risk. (F)Schematic showing the amino acid substitutions encoded by the NKX3.1 polymorphisms, Arginine 52 to Cysteine (R52C) and Threonine 164 to Alanine (T164A), or the control, Arginine 52 to Alanine (R52A). Also shown is a summary of the DNA binding, mitochondrial localization and ROS response. High level activity is indicated by (+) and no/low level activity by (−). (G-I) Analyses of RWPE1 cells expressing NKX3.1 (wildtype), T164A, R52C, or R52A treated with vehicle or paraquat (100μM) for 24 hours. (G) Western blot analyses following isolation of nuclear or mitochondrial fractions. (H and I) Quantification of mean ROS intensity, as detected by DHE (H) or MitoPY1 (I) to measure general cellular or mitochondrial ROS, respectively. (J-P) Analyses of LNCaP cells expressing shNKX3.1 (or a control shRNA) alone or with exogenous NKX3.1, T164A, or R52C. Cells were treated with vehicle or paraquat (100μM) for 24 hours. (J) Western blot analyses following isolation of nuclear or mitochondrial fractions. (K and L) Quantification of mean ROS intensity, as detected by DHE (K) or MitoSOX (L). (M) Quantification of cellular proliferation as detected by MTT absorbance. (N) Representative images of colony formation assays visualized by staining for crystal violet. (O) Quantification of colony number. (P) Quantification of Matrigel invasion assays. Unless otherwise indicated, shown are representative data from 2-3 independent experiments, each done in triplicate (9 independent samples/group) showing mean ± SD; P values were calculated using two-sample unpaired Welch t-test. ns, not significant. See also Supplementary Figures S6 and S7 and Supplementary Tables S3 and S5.

Figure 3.. HSPA9 imports NKX3.1 to mitochondria to protect from oxidative stress and suppress tumorigenicity

(A) Strategy to isolate NKX3.1-associated proteins in mitochondria. RWPE1 cells expressing a FLAG-HA-tagged NKX3.1 were treated with paraquat (100μM) for 24 hours followed by biochemical fractionation to isolate mitochondria. NKX3.1-interacting proteins were isolated by immunoprecipitation and identified by mass spectrometry (see Methods). (B) Western blot analysis following co-immunoprecipitation of mitochondria from paraquat-treated RWPE1 cells expressing FLAG-HA-tagged NKX3.1 (or a control vector). NKX3.1 protein complexes were isolated using anti-FLAG followed by Western blot detection with the indicated antibodies. (C) Western blot analysis following co-immunoprecipitation of NKX3.1, T164A, and R52C to detect HSPA9. RWPE1 cells expressing FLAG-HA-tagged proteins (or a control vector) were treated with paraquat (100μM) or vehicle for 24 hours. In panels B and C, input represents 10% of the total protein used for immunoprecipitation; IP, immunoprecipitated proteins. (D, E) RWPE1 cells expressing NKX3.1 (or a control vector) were infected with an shRNA for HSPA9 (shHSPA9) and treated with paraquat (100μM) or vehicle for 24 hours. (D) Western blot analysis of nuclear and mitochondrial fractions. (E) Quantification of mean ROS intensity as detected by DHE production. (F-M) Analysis of LNCaP cells expressing shRNAs for HSPA9 (shHSPA9) and/or NKX3.1 (shNKX3.1) or the control (shControl), as indicated. Cells were treated with paraquat (100μM) or vehicle for 24 hours. (F) Western blot analysis of nuclear and mitochondrial fractions. (G) Confocal images of LNCaP cells co-stained with anti-NKX3.1 (red), anti-HSPA9 (cyan), and anti-ATPB to visualize mitochondria (green); nuclei were visualized by labeling with DAPI (blue). Scale bars represent 25μm. (H and I) Quantification of mean ROS intensity as detected by DHE (H) or MitoSOX (I). (J) Quantification of cellular proliferation as detected by MTT absorbance. (K) Representative images of colony formation assays visualized by staining with crystal violet. (L) Quantification of number of colonies. (M) Quantification of Matrigel invasion assays. Unless otherwise indicated, shown are representative data from 2-3 independent experiments, each done in triplicate (9 independent samples/group). Data are expressed mean ± SD; P values were calculated using a two-sample unpaired Welch t-test. ns, not significant. See also Supplementary Figures S8 and S9, Supplementary Tables S3, S4, and S6, and Dataset 5.

Figure 4.. NKX3.1 regulates expression of mitochondrial-encoded electron transport chain genes

(A) Schematic diagram of OXPHOS complexes of the electron transport chain (ETC) located in the inner mitochondrial membrane. Shown are sites of NAD+ (Complex I) and ATP (Complex V) production as well as the major sources of ROS (Complex I and III). Complexes I, III, and IV act as proton (H⁺) pumps to facilitate electron transport, whereas Complex V moves H⁺ from the intermembrane space to the mitochondrial matrix to convert ADP to ATP. Free electrons lead to ROS accumulation in the form of superoxide (O⁻₂) or hydrogen peroxide (H₂O₂). (B-F) Analyses of RWPE1 cells expressing NKX3.1, T164A, or R52C (or the control vector) treated with vehicle or paraquat (100μM) for 24 hours. (B,C) Heat map representations of nuclear-encoded and/or mitochondrial-encoded (MT-encoded) ETC genes based on RNA sequencing analyses. (B) Differential expression of MT-encoded ETC genes in cells expressing NKX3.1, T164A, or R52C. (C) Differential expression of nuclear-encoded and MT-encoded ETC genes in NKX3.1-expressing cells; MT-encoded genes are shown by arrows. (D) Quantitative real-time PCR analysis showing relative expression of MT-encoded ETC genes. Data are expressed as relative mRNA levels (relative to 18s rRNA expression) showing the mean ± SD; (E-F) NKX3.1 binding to the D-Loop in the mitochondrial genome. (E) Schematic representation of the D-loop, indicating the location of the NKX3.1 consensus DNA binding site (TAAGTA) and the positions of two independent primer sets (primer sets A and B) used for Chromatin immunoprecipitation-quantitative PCR (ChIP-qPCR) (arrows). LSP, light strand promoter position. HSP1, heavy strand promoter 1 position. HSP2, heavy strand promoter 2 position. (F) ChIP-qPCR of NKX3.1 binding to the D-loop in RWPE1 cells expressing exogenous NKX3.1, T164A, or R52C. Cells were treated with paraquat (100μM) or vehicle, and ChIP was done using an antibody against human NKX3.1. Data are expressed as relative enrichment of NKX3.1 binding showing the mean ± SD. Unless otherwise indicated, shown are representative data from 3 independent experiments, each done in triplicate (9 independent samples/group). P values were calculated using a two-sample unpaired Welch t-test. ns, not significant. See also Supplementary Figures S10-S12 and Dataset 2.

Figure 5.. NKX3.1 regulates OXPHOS activity and mitochondrial respiration

(A-D) Analyses of RWPE1 cells expressing NKX3.1, T164A, or R52C treated with paraquat (100μM) or vehicle for 24 hours. (A) Quantification of activity of OXPHOS complexes I, II, III, IV and V from mitochondria isolated from paraquat-treated cells (n=5 independent samples/group). (B) Western blot analyses of nuclear and total mitochondrial fractions to assess expression levels of the mitochondrial-encoded (mt) OXPHOS complex proteins as indicated. (C) NADH/NAD⁺ ratio. (D) Relative ATP levels. (E-H) Seahorse analyses of mitochondrial respiration and glycolysis in NKX3.1-expressing (or control) RWPE1 and NKX3.1-silenced (or control) LNCaP cells. (E and F) Oxygen consumption rate (OCR) analyses. The rates of basal-, ATP-linked, maximal, and reserve respiration were quantified by normalization of OCR level to the total protein optical density (OD) values. (G and H) Extracellular acidification rate (ECAR) analyses. The rates of glycolysis, glycolytic capacity and reserve were quantified by normalization of ECAR level to the total protein OD values. Unless otherwise indicated, shown are data from 3 independent experiments, each done in triplicate (9 independent samples/group); P values were calculated using two-sample unpaired Welch t-test. ns, not significant. See also Supplementary Figure S13.

Figure 6.. Association of NKX3.1 expression levels with mitochondrial function and clinical outcome in human prostate cancer

(A-H) Analyses of primary human organotypic cultures. (A) Strategy. Primary human prostate cancer tissues were obtained directly from surgery and cultured in vitro in media alone (vehicle) or media containing paraquat (100μM) for 24 hours. (B) Representative images showing histology (H&E) and NKX3.1 immunostaining. Scale bars represent 25μm. (C) Representative confocal images co-stained for NKX3.1 (red) and ATPB (green); nuclei were visualized with DAPI (blue). Scale bars represent 25μm (left and center) or 200μm (right). (D) ChIP-qPCR of NKX3.1 binding to the D-loop of the mitochondrial genome in human prostate organotypic cultures. ChIP was done using an antibody that recognizes human NKX3.1. Data are expressed as relative enrichment of NKX3.1 binding showing the mean ± SD. (E-H) Analyses of human prostate organotypic cultures having “high” versus “low” levels of NKX3.1 expression. (E) Heat map showing NKX3.1 expression levels determined by RT-qPCR analyses. (F) RT-qPCR analysis of mitochondrial-encoded ETC genes in paraquat- versus vehicle-treated organotypic cultures. Data are expressed as relative mRNA levels (relative to 18s rRNA expression) showing the mean ± SD. (G) NADH/NAD⁺ ratio. (H) ATP levels. Panels F-H shows representative data from 3 independent experiments, each done in triplicate (9 independent samples/group). P values were calculated using two-sample unpaired Welch t-test. ns, not significant. (I) Biological pathway-based GSEA using a gene signature comparing “high” versus “low” NKX3.1 mRNA expression from the Taylor cohort to query pathways from the C2 pathway collection (see Methods). “High” corresponds to the top 25% of patients with highest levels of NKX3.1 expression (n=32) and “Low” corresponds to the top 25% of patients with the lowest levels of NKX3.1 expression (n=32). Normalized Enrichment Score (NES) and P values were calculated based on 1000 permutations. (J) Kaplan-Meier survival analysis showing association of NKX3.1 and mitochondrial-encoded ETC (mito) expression levels with biochemical recurrence (BCR) estimated survival probability based on the TCGA cohort. Analyses compare patients with ‘high’ or ‘low’ expression levels of NKX3.1 and “high” or “low” expression levels of the combined 13 mitochondrial ETC genes (mito); P values were estimated using a log-rank test. See also Supplementary Figure S14, Supplementary Tables S7 and S8, and Dataset 6.

Figure 7.. Association of non-nuclear NKX3.1 protein with clinical outcome in human prostate cancer

(A) Strategy. NKX3.1 protein expression levels were examined on two independent TMAs. TMA1 is comprised of 194 patients with Gleason score 6 or 7 tumors; TMA2 is comprised of 118 patients with Gleason score 5-9 tumors. Immunostaining of NKX3.1 was graded based on nuclear or non-nuclear expression and corresponding to “high”, “low”, or “negative” expression. (B) Representative cases based on TMA2 showing nuclear or non-nuclear NKX3.1 immunostaining and examples of “high”, “low”, or “negative” expression. Scale bars represent 50μm. (C) Representative confocal images of human prostate tissues co-stained for NKX3.1 (red) and ATPB (green); nuclei were visualized with DAPI (blue). Scale bars represent 25μm. (D) Representative human prostate tumors showing examples of “high” or “low” NKX3.1. Adjacent sections were stained for HSPA9 and 4HNE (marker of oxidative damage). (E) Analyses of TMA2, showing the association of “high” or “low/no” expression levels of total, nuclear, or non-nuclear NKX3.1 protein relative to pre-operative PSA levels (total n of patients = 82). P values were calculated using two-sample unpaired Welch t-test. ns, not significant. (F, G) Kaplan-Meier survival analyses showing association of total, nuclear, or non-nuclear NKX3.1 protein expression levels with biochemical recurrence (BCR)-free estimated survival probability for TMA1 (n=112) (Panel F) and overall survival probability for TMA2 (n=90) (Panel G); P values were estimated using a log-rank test. See also Supplementary Figure S14 and Supplementary Table S8.