Abstract

   Targeted α-particle therapy is a promising option for patients with
   malignant pheochromocytoma. Recent observations regarding
   meta-^211At-astato-benzylguanidine (^211At-MABG) in a pheochromocytoma
   mouse model showed a strong anti-tumor effect, though the molecular
   mechanism remains elusive. Here, we present the first comprehensive
   RNA-sequencing (RNA-seq) data for pheochromocytoma cells based on in
   vitro ^211At-MABG administration experiments. Key genes and pathways in
   the tumor α-particle radiation response are also examined to obtain
   potential response biomarkers.

   Methods: We evaluated genome-wide transcriptional alterations in the
   rat pheochromocytoma cell line PC12 at 3, 6, and 12 h after ^211At-MABG
   treatment; a control experiment using ^60Co γ-ray irradiation was
   carried out to highlight ^211At-MABG-specific gene expression. For
   comparisons, 10% and 80% iso-survival doses (0.8 and 0.1 kBq/mL for
   ^211At-MABG and 10 and 1 Gy for ^60Co γ-rays) were used.

   Results: Enrichment analysis of differentially expressed genes (DEGs)
   and analysis of the gene expression profiles of cell cycle checkpoints
   revealed similar modes of cell death via the p53-p21 signaling pathway
   after ^211At-MABG treatment and γ-ray irradiation. The top list of
   ranked DEGs demonstrated the expression of key genes on the decrease in
   the survival following ^211At-MABG exposure, and four potential genes
   (Mien1, Otub1, Vdac1 and Vegfa genes) of ^211At-MABG therapy. Western
   blot analysis indicated increased expression of TSPO in
   ^211At-MABG-treated cells, suggesting its potential as a PET imaging
   probe.

   Conclusion: Comprehensive RNA-seq revealed contrasting cellular
   responses to γ-ray and α-particle therapy, leading to the
   identification of four potential candidate genes that may serve as
   molecular imaging and ^211At-MABG therapy targets.

   Keywords: α-particle, meta-^211At-astato-benzylguanidine,
   RNA-sequencing, pheochromocytoma, PET imaging, radionuclide therapy

Introduction

   Pheochromocytomas (PCCs) are rare neuroendocrine tumors with malignant
   progression that account for approximately 30% of extra-adrenal PCCs
   [39]^1. Some malignant PCCs exhibit systemic metastasis, and clinical
   improvement with β-emitting meta-^131I- iodo-benzylguanidine
   (^131I-MIBG) constitutes a stage of partial remission in metastatic PCC
   [40]^2. Recently, we reported strong anti-tumor effects of α-emitting
   meta-^211At-astato-benzylguanidine (^211At-MABG) in a PCC mouse model,
   suggesting a potential option for targeted α therapy (TAT) for patients
   with malignant PCC [41]^3. Nonetheless, very little is known about the
   molecular mechanism of the therapeutic effects of ^211At-MABG compared
   with photon-radiation therapy [42]^4. To avoid unintended consequences,
   a better understanding of the molecular basis of such therapeutic
   effects is necessary for the clinical implementation of ^211At-MABG
   therapy in malignant PCC.

   Theranostics, which combines specific targeted therapy and specific
   targeted diagnosis, is important for TAT, especially with regard to
   diagnosis. In vivo localization is one of main components of
   theranostic diagnosis, which includes Cherenkov luminescence imaging of
   ^225Ac-labeled compounds [43]^5 and a Compton camera for ^211At-labeled
   compounds [44]^6. Another main component is the real-time dynamic
   imaging proposed by Paulmurugan et al. [45]^7. Overall, understanding
   or assessing therapeutic tumor responses, e.g., through inflammation
   biomarkers, is crucial [46]^8; although both components are required
   for TAT, few studies have focused on the latter. Accordingly, in this
   study, we explored therapeutic response biomarkers.

   Post-genome sequencing technologies have rapidly developed in the last
   ten years, and next-generation sequencing (NGS) is a powerful tool for
   unraveling the molecular basis of rare genetic diseases [47]^9.
   Furthermore, RNA-sequencing (RNA- seq) using NGS provides a more
   precise transcription profile compared to other methods [48]^10, and
   this approach may contribute to the search of TAT- response biomarkers.

   Prior to the advent of RNA-seq, some in vitro TAT studies utilized
   microarray technology to measure gene expression and identify
   differentially expressed genes (DEGs) participating in DNA repair, cell
   cycle checkpoints, and apoptosis [49]^11^, [50]^12^, [51]^13, as shown
   by photon radiation therapy [52]^4. For example, Seidl et al. reported
   novel gene regulation of TAT with high significance, identifying
   COL4A2, NEDD9, and C3 genes as being continuously upregulated; however,
   their specific functions in the response to high linear-energy-transfer
   ionizing radiation (IR) remain unknown [53]^11. Comprehensive RNA-seq
   analysis can now provide a wide dynamic range and great statistical
   precision with an increasing number of time points and biological
   replicates, thus allowing precise identification of key genes of TAT
   [54]^14.

   In this study, we examined gene expression profiles via RNA-seq in rat
   PCC cells to elucidate the molecular mechanism of ^211At-MABG
   therapeutic effects compared with those of photon (γ-ray) irradiation
   under the expected conditions in line with TAT and the conventional
   radiotherapy. We further explored key genes in the tumor response to
   radiation as well as potential molecular therapeutic biomarkers for
   malignant PCC.

Materials and Methods

Cell culture

   PC12, a rat pheochromocytoma cell line, is a representative cell line
   for malignant PCC [55]^15 with long history of being the model for
   nuclear medicine studies, including the contribution to the preclinical
   study of MIBG therapy [56]^16. PC12 was purchased from Japanese
   Collection of Research Bioresources (IFO50278, Osaka, Japan), and
   cultured as previously reported [57]^3 ([58]Supplementary materials and
   methods).

^211At-MABG treatment, γ-ray irradiation and dose estimation

   ^211At-MABG was synthesized as previously described [59]^3. The
   radioactivity of ^211At (T[1/2] = 7.2 h) was measured from γ-rays
   emitted in ^211At decay using a high-purity germanium detector. We
   estimated the absorbed dose of ^211At-MABG-treated cells using a
   published method [60]^17, with some modifications ([61]Supplementary
   materials and methods), as based on cellular uptake and release
   experiments, a parameter of energy per decay for ^211At [62]^18 and a
   real-coded genetic algorithm to estimate parameters [63]^19.
   Stable-iodine labeled MIBG was used for the nonradioactive experiments
   of ^211At-MABG (MIBG-control; [64]Supplementary materials and methods).
   The ^60Co source at the Takasaki food irradiation facility was used for
   γ-ray irradiation, and the dose-rate distribution was routinely
   monitored using polymer-alanine dosimeters (Hitachi Cable, Ltd., Tokyo,
   Japan). The absorbed dose of γ-ray irradiated cells, assumed to be a
   water equivalent, was interpolated from routine monitoring data. PC12
   cells were exposed to ^211At-MABG at concentrations of 0, 0.2, 0.6, 2.0
   and 6.0 kBq/mL for 1 day or were irradiated with ^60Co γ-rays at doses
   of 0, 0.1, 0.3, 1, 3 and 10 Gy for 12 min time periods. Figure [65]1
   shows the experimental design for ^211At-MABG treatment and ^60Co
   γ-rays irradiation. Cell survival assays were carried out as previously
   reported ([66]Supplementary materials and methods).

Figure 1.

   [67]Figure 1
   [68]Open in a new tab

   Experimental design for ^211At-MABG treatment and ^60Co γ-rays
   irradiation. Comparative RNA-seq analysis between control,
   γ-ray-irradiated and ^211At-MABG-treated samples was performed at 3, 6
   and 12 h. MIBG-control additional experiment was also carried out in
   the same time course. Cell cycle distribution was measured at 24 h, and
   a cell survival assay was performed after 2 weeks of incubation.

RNA extraction and sequencing

   A sample of 10^6 harvested cells was resuspended in TRIzol (Thermo
   Fisher Scientific, Waltham, MA, USA), and RNA was extracted using the
   Direct-zol RNA kit (Zymo Research Corp, Orange, CA, USA) according to
   the manufacturer's protocol. A sequencing library was prepared using
   the NEBNext, Ultra RNA Library Prep Kit (New England Biolabs, Ipswich,
   MA). Multiplex sequencing was performed with the High Output Mode
   regents of the NextSeq 500 instrument (Illumina Inc., San Diego, CA)
   using 75 cycles for control and γ-ray-irradiated samples and 300 cycles
   for ^211At-MABG treated samples. These sequencing data sets were
   deposited (and are available) at the DNA Data Bank of Japan (DDBJ:
   [69]http://www.ddbj.nig.ac.jp/) Sequence Read Archive under Accession
   number DRA007102 and DRA007735.

Differential expression analysis

   Sequencing reads were aligned to the rat Ensembl Rnor 6.0 reference
   genome obtained from Ensembl using TopHat2 v2.0.14 with a default
   setting [70]^20. The aligned reads were assembled using Cufflinks
   v2.2.1 [71]^21 with an annotation file (Rnor 6.0.83), along with
   estimation of the abundance of each transcript. Expression levels were
   calculated as fragments per kilobase of exon per million fragments
   mapped (FPKM). Assemblies of three replicates were combined using
   Cuffmerge v1.0.0. Differential expression analysis of each combination
   was performed using Cuffdiff v2.2.1 [72]^22.

   For pathway enrichment analysis, reference genes were first matched to
   KEGG (Kyoto Encyclopedia of Genes and Genomes) orthology using the KEGG
   KAAS web service [73]^23, and enrichment of pathways for each time
   point was assessed using a custom Perl script in the G-language Genome
   Analysis Environment [74]^24, utilizing Fisher's exact test of the R
   package ([75]https://www.r-project.org) with Bonferroni correction of
   p-values for multiple testing. For both differential gene expression
   and pathway enrichment analysis, a false discovery rate (FDR) threshold
   of 0.05 was used. Disease-related pathways with a pathway map ID
   greater than 5000 were omitted.

Configuration of the gene regulatory network and representative DEGs

   A gene regulatory network was obtained from the KEGG pathway map
   [76]^23. Genes involved in the regulatory network were screened
   according to the following conditions: 1. at least twofold highly
   expressed genes in the treatment; 2. at least FPKM > 1 for all
   conditions; 3. overrepresentation in at least one of twelve conditions
   (γ-ray irradiation, ^211At-MABG treatment, two survival rates, and
   three time points) compared with the corresponding controls; 4.
   presence of the canonical gene name.

   Using similar conditions, except for the 3rd, representative genes with
   significant changes in expression were screened between the γ-ray and
   ^211At-MABG treatments. The 3^rd condition was modified with a
   comparison of combination treatments: e.g., for representative DEGs of
   ^211At-MABG treated cells between 80% and 10% survival doses at each
   time point, we used the condition of DEGs overrepresented in at least
   one of the two conditions (including the two survival rates).

Flow cytometry and western blot analysis

   Cell cycle distributions among cells treated with ^211At-MABG (0.6
   kBq/mL) for 24 h and 24 h post-irradiated with γ-ray (10 Gy) were
   analyzed using flow cytometry. Cells were collected and fixed in 70%
   ethanol overnight, treated with RNase for 20 min before the addition of
   5 μg/mL of propidium iodide and analyzed by flow cytometry (EC-800,
   Sony, Tokyo, Japan). Western blot analysis of cells treated with
   ^211At-MABG (0.8 kBq/mL) for 6, 12 and 24 h was performed as previously
   described ([77]Supplementary materials and methods) [78]^25.

Statistical analysis

   The computer program (ORIGIN, MicroCal Software, Inc., MA, USA) was
   used for non-linear curve fitting.

Results

Iso-survival dose and experimental design for ^211At-MABG treatment and ^60Co
γ-ray irradiation

   To compare the biological effects of two types of radiation, we first
   evaluated the 10% and 80% iso-survival doses of ^60Co γ-ray irradiation
   and ^211At-MABG treatment for PC12 cells. For this study, we used the
   acute irradiation for ^60Co γ-rays to imitate conventional radiation
   therapy.

   Cells irradiated with γ-rays exhibited a log- linear relationship
   between the surviving fraction (SF) and the absorbed dose (Figure
   [79]2), with 10% and 80% iso-survival doses of γ 10 Gy (SF = 10.8%) and
   1 Gy (SF = 80.1%), respectively. The absorbed dose of ^211At- MABG by
   treated cells was estimated using a published method [80]^19, with some
   modifications ([81]Supplementary materials and methods). When cells
   were exposed to ^211At-MABG at the initial concentration of 1.0 kBq/mL,
   the absorbed dose was estimated to be 4.31 Gy. Using this estimation,
   the relationship between the SFs of 0, 0.2, 0.6, 2.0 and 6.0 kBq/mL and
   the absorbed dose displayed a sigmoidal curve (Figure [82]2). The 10%
   and 80% iso-survival doses of ^211At- MABG were set to 0.8 kBq/mL (3.5
   Gy; SF = 9.8%) and 0.1 kBq/mL (0.4 Gy; SF = 82%), respectiveely, and we
   designed the time-course of RNA-seq analysis accordingly (Figure
   [83]1). In general, IR- induced DNA damage causes an increase or
   decrease in gene expression several hours post-irradiation. Indeed,
   ^60Co γ-rays and α-particles emitted from ^211At induced significant
   alterations in gene expression several hours after treatment [84]^14.
   Thus, we used three time points for RNA-seq analysis at 3, 6 and 12 h
   after ^60Co γ-ray irradiation and the start of ^211At-MABG exposure.

Figure 2.

   [85]Figure 2
   [86]Open in a new tab

   Iso-survival dose for ^211At-MABG treatment and ^60Co γ-rays
   irradiation. Surviving fraction (SF) after ^211At-MABG treatment and
   ^60Co γ-rays irradiation in PC12 cells. The SF of ^211At-MABG treatment
   demonstrated a sigmoidal curve (^211At-MABG Fit)—that is, SF (%) = 1.3
   + (100.0 - 1.3) / (1.0 + (Dose (Gy) / 0.97)^1.9) (the correlation
   coefficient square, R^2 = 1.00). The SF of ^60Co γ-ray irradiation
   showed a log-linear relationship (γFit)—that is, SF (%) = 100.0 exp (
   -0.22 * Dose (Gy) ) (R^2 = 0.98). Thin lines indicate 10% and 80%
   survival.

Comprehensive RNA-seq analysis of ^211At-MABG-treated and ^60Co
γ-ray-irradiated PC12 cells

   We employed a transcriptomic approach to observe changes in gene
   expression in PC12 cells at iso-survival doses of 10% and 80%, and a
   minimum of 7 million reads were sequenced for each condition in three
   replicates (Figure [87]3). Overall transcriptome profiles were
   consistent among the replicates and strikingly distinct among the three
   treatment conditions (control, γ-ray irradiation, ^211At-MABG
   treatment), essentially regardless of the survival rate or time point.
   Although γ-ray-irradiated samples shared a proportion of gene clusters
   showing similar expression levels as the control,^ 211At-MABG treatment
   resulted in vastly different profiles, with the early (3 h) response to
   weak irradiation (80% survival) exhibiting a very weak correlation
   (approximately 0.2 in Pearson correlation) with the 3-h control
   conditions (Figure [88]S3). These differential expression patterns were
   further supported by the number of DEGs, as summarized in Figure
   [89]S4. The numbers of DEGs were highly consistent among the different
   time points, with the number of DEGs at 80% survival after ^211At-MABG
   treatment (5,881, 5,198 and 5,304 genes at 3, 6 and 12 h, respectively)
   all surpassing those of 10% survival after γ-ray irradiation (4,438,
   3,812 and 4,414 genes at 3, 6 and 12 h, respectively). MIBG-control
   experiments did not show the same transcriptome profiles treated with
   ^211At-MABG and ^60Co γ-rays (Figure [90]S5), and showed very limited
   number of DEGs, where the maximum number of DEGs exceeding fold change
   > 2 being 70 (Figure [91]S6). Significant changes in gene expression at
   80% survival after γ-ray irradiation (1,974, 1,731 and 2,590 genes at
   3, 6 and 12 h, respectively) were largely fewer than half under all
   other conditions. Therefore, the intracellular response to these two
   conditions appears to be contrasting, presumably providing an
   explanation for the different tumor responses to X/γ-ray and TAT
   therapies. We thus focus on the difference in responses, in other
   words, the more than 5,000 DEGs between the two irradiation conditions.

Figure 3.

   [92]Figure 3
   [93]Open in a new tab

   Heatmap clustering of all RNA-seq data used in this analysis. The
   expression level of all genes with TPM (Transcript Per Million) > 5 in
   any condition is normalized relative to the median expression level for
   each gene. The expression level was calculated using kallisto software
   (v.0.42.4). Regardless of the time point or survival rate, the three
   conditions (control, γ-ray irradiation, ^211At-MABG treatment) showed
   distinct expression patterns, with ^211At-MABG treatment showing
   greater variability.

   Enrichment in biological processes in DEGs was next examined using KEGG
   pathway enrichment analysis (Figure [94]4). γ-Ray irradiation
   predominantly resulted in under-representation of DNA damage repair
   (base excision repair, nucleotide excision repair, mismatch repair, and
   homologous recombination) and cell cycle-related (DNA replication, cell
   cycle) pathways, suggesting cell cycle arrest and stalling of DNA
   replication. In contrast, ^211At-MABG treatment did not result in these
   under- representations, except for 6 or 12 h of 10% survival; however,
   over-representation of protein turnover (ribosome biogenesis, RNA
   transport, ubiquitin mediated proteolysis, and protein processing in
   endoplasmic reticulum) was observed. Only the p53 signaling pathway was
   over-represented in cells treated with both γ-ray irradiation and
   ^211At-MABG.

Figure 4.

   [95]Figure 4
   [96]Open in a new tab

   Enrichment of KEGG pathways in DEGs at each condition. KEGG (Kyoto
   Encyclopedia of Genes and Genomes) enrichment for differentially
   expressed genes (DEGs) was computed using Fisher's exact test with
   Bonferroni correction of p-values, where the corrected p-value < 0.05
   is shown. KEGG pathways for human diseases are omitted. Red and blue
   squares represent positive and negative enrichment, respectively.
   γ-Ray-irradiated and ^211At-MABG-treated samples show contrasting
   enrichment of KEGG pathways.

Cell cycle checkpoints and the p53-p21 pathway

   The pathway enrichment analysis indicated cell cycle arrest, stalling
   of DNA replication and activation of p53 signaling. To understand how
   γ-ray irradiation and ^211At-MABG treatment regulate the cell cycle, we
   screened well-known genes that participate in DNA damage-dependent G1/S
   and G2/M checkpoints using the present configuration (see a description
   in the Materials and methods). The relationships among Tp53, Mdm2,
   Chek2, Cdkn1a, Skp2, Cdk2, Ccne1,2, Rb1, Gadd45a,g, Cdc25b, Plk1, Cdk1
   and Ccnb1 transcripts are illustrated in Figure [97]5A. Despite
   differences in the gene expression pattern for the early response and
   delayed induction between both treatments (Figure [98]S7), an overview
   of the selected gene expression levels resulted in a similar expression
   profile (Figure [99]5A). The expression levels of Mdm2, Cdkn1a,
   Gadd45a,g and Rb1 genes increased, whereas those of the other genes
   decreased (Figure [100]5A), suggesting activation of G1/S and G2/M
   checkpoints via the p53 pathway. Flow cytometry analysis supported the
   transcriptomic results, demonstrating a decrease and increase in the
   ratio of S- and G2/M-phase cells, respectively (Figure [101]5B). These
   results show that the p53 pathway is an important cell cycle regulator
   in γ-ray irradiation and ^211At-MABG treatment. Nonetheless,
   transcriptional activation of the Cdkn1a gene is also regulated via
   p53-independent pathways [102]^26 (Figure [103]5C). In addition, we
   found reduced expression of suppressor APC-MYC pathway components and
   increased Klf6 expression to be associated with transcriptional
   regulation of CDKN1A (Figure [104]5C-D). Taken together, our data
   suggest that the p53 signaling pathway via DNA damage and associated
   signals regulate the PC12 cell cycle after γ-ray irradiation and
   ^211At-MABG treatment.

Figure 5.

   [105]Figure 5
   [106]Open in a new tab

   Cell cycle checkpoints, cell cycle distribution and the p53-p21
   pathway. (A) The gene regulatory network for cell cycle checkpoints was
   configured with reference to the KEGG pathway map [107]^17. Red and
   blue ellipses represent the increase or decrease in gene expression
   (vs. control) after both ^60Co γ-ray irradiation and ^211At-MABG
   treatment, respectively. Both treatments showed similar gene expression
   profiles. (B) Cells per phase in % of 10% iso-survival experiments
   (left and right panels for γ-ray irradiation and ^211At-MABG treatment,
   respectively). The error bars indicate the standard error of mean. (C)
   Cdkn1a transcriptional circuitry [108]^19. Red and blue ellipses
   represent the increase or decrease in gene expression (vs. control),
   respectively. (D) Gene expression over time of Myc, Apc, Klf6 and Tgfb1
   genes. The error bars represent standard deviation.

Representative DEGs of ^211At-MABG-treated cells between 80% and 10% survival
doses

   Analysis of cell cycle checkpoints demonstrated a similar response to
   γ-ray irradiation and ^211At-MABG treatment, yet the DEGs associated
   with ^211At-MABG therapeutic effects were unclear. Therefore, we
   investigated representative DEGs that contribute to a decrease in
   survival after ^211At-MABG treatment.

   The number of representative DEGs between 80% and 10% survival
   increased in a time-dependent manner (11 genes at 3 h post-treatment to
   60 genes at 12 h). Table [109]1 classifies the representative DEGs into
   5 categories at each time point. The top-ranking genes were Nudt6
   (7.9-fold increase), Eda2r (5.3-fold increase) and Inmt (5.3-fold
   increase) at 3, 6 and 12 h post-^211At-MABG treatment, respectively.
   Nudt6 is suggested to play a role in cell proliferation [110]^27, and
   Eda2r is reported to induce apoptosis and prevent cell adhesion through
   the p53-regulated anoikis pathway [111]^28. Inmt codes for an
   indolethylamine-N-methyltransferase (INMT), which produces
   N,N-dimethyltryptamine (DMT) [112]^29; although the role of endogenous
   DMT has not yet been established, exogenous DMT acts as an inhibitor of
   peripheral monoamine oxidase [113]^30. Representative DEGs found at all
   time points included Gdf15, Fam212b, Cdkn1a, Enc1 and Tp53inp1 (Table
   [114]2), all of which are target genes of activated p53 [115]^31. These
   DEGs provide a straightforward description of the state of PC12 cells
   incorporating ^211At-MABG, i.e., cell cycle regulation and cell death
   induction.

Table 1.

   ^211At-MABG induces a time-dependent shift in representative DEGs
   between 80% and 10% survival experiments.
   3 h post-treatment (11 genes)^*1 6 h post-treatment (52 genes)^*1 12 h
   post-treatment (60 genes)^*1
   Gene name LogFC^*2 Function Gene name LogFC^*2 Function Gene name
   LogFC^*2 Function
   Cell cycle Nudt6 +2.976 (1) Cell proliferation Ptprv +1.889 (6) G1/S
   checkpoint Ptprv +2.022 (5) G1/S checkpoint
   Ccng1 +1.412 (12) G2/M checkpoint Ccng1 +1.412 (25) G2/M checkpoint
   Plk2 +1.055 (44) Spindle checkpoint Plk2 +1.182 (43) Spindle checkpoint
   Cell death Eda2r +2.403 (1) Apoptosis and prevent ion of cell adhesion
   Eda2r +2.006 (6) Apoptosis and prevention of cell adhesion
   Aen +1.354 (16) Apoptosis Atg9b +1.991 (7) Autophagy
   Btg2 +1.351 (18) Apoptosis Gsdmd +1.669 (13) Programmed cell death
   DNA repair Aplf +1.381 (5) DNA repair Ier5 +1.201 (28) DNA repair Mgmt
   +1.290 (31) DNA repair
   Fam175a +1.018 (11) DNA repair with BRCA1 Rnf169 +1.028 (48) Homologous
   recombination Xrcc1 +1.044 (54) Base excision repair
   Metastasis Adam8 +1.882 (7) Invasion Adam8 +2.164 (3) Invasion
   Jam3 +1.177 (30) Adhesion Mmp2 +1.851 (8) Metastasis
   Others Inmt +1.631 (9) Methyltransferase Inmt +2.403 (1)
   Methyltransferase
   Trim7 +1.315 (21) Proliferation, Apoptosis Lrrc25 +1.682 (11)
   Inflammation,
   Autophagy
   [116]Open in a new tab

   ^*1 Number of DEGs. ^*2 The number in parentheses indicates the rank

Table 2.

   ^211At-MABG induced representative DEGs for decreased survival at all
   time points.
   All time points (5 genes)^*1
   Gene name LogFC^*2, *3 Function
   Cell cycle Cdkn1a +1.680 (3) G1/G2 checkpoint
   Cell death Enc1 +1.657 (4) Negatively regulates autophagy, redox
   homeostasis
   Tp53inp1 +1.320 (5) Apoptosis, Autophagy
   Metastasis Gdf15 +1.782 (1) Epithelial mesenchymal transition
   Others Fam212b +1.503 (2) Unknown
   [117]Open in a new tab

   ^*1 Number of DEGs. ^*2 The number in parentheses indicates the rank.
   ^*3 Value at 12 h post-treatment.

   Furthermore, representative DEGs in PC12 cells at each time point
   exhibited varying gene expression profiles. At 3 h post-treatment, Aplf
   and Fam175a genes were observed, with Fam175a suggesting the
   involvement of homologous recombination (HR). DNA-repair related genes
   were continuously expressed after 3 h post-treatment (Ier5, Rnf169,
   Mgmt and Xrcc1of Table [118]1); various genes were represented at 6 and
   12 h post-treatment. In the cell cycle category, Ptprv for the G1/S
   checkpoint, Ccng1 for the G2/M checkpoint and Plk2 for the spindle
   checkpoint were highly represented, in agreement with the results shown
   in Figure [119]5A-B. Regarding the category of cell death,
   apoptosis-related genes (Aen, Btg2 and Trim7) were expressed from 6 h
   post-treatment, whereas autophagy- and other-related genes (Atg9b,
   Gsdmd and Lrrc25) ranked in the top 30 at 12 h post-treatment.
   Moreover, ^211At-MABG treatment induced metastasis-related gene
   expression (Adam8, Jam3 and Mmp2). Therefore, representative DEGs of
   ^211At-MABG-treated cells between 80% and 10% survival doses
   effectively highlight the expression of key genes associated with
   ^211At-MABG therapeutic and anti-therapeutic effects.

Potential therapeutic response biomarkers of ^211At-MABG

   To investigate the different molecular mechanisms of ^211At-MABG
   treatment compared with photon (γ-ray) irradiation, we further screened
   for representative genes displaying changes in expression between γ-ray
   and ^211At-MABG treatments. Representative DEG analysis revealed in 161
   genes (Table [120]S1), among which 26 showed a log fold change greater
   than 2 and 87 were highly abundant, with a maximum FPKM > 100. Thus,
   only 20 genes satisfied both conditions, with 10 annotated in KEGG
   pathways. We selected four of these genes, Mien1, Otub1, Vdac1 and
   Vegfa, based on their possible contribution to the ^211At-MABG
   treatment response. Overall, their maximum log fold-changes in gene
   expression (15.5, 8.4, 9.5 and 6.9 vs. γ-ray irradiation; Figure
   [121]6A) and continuous high level of expression demonstrate the
   ^211At-MABG-specific response.

Figure 6.

   [122]Figure 6
   [123]Open in a new tab

   Potential molecular biomarkers for ^211At-MABG-therapy. (A) Gene
   expression according to the FPKM of Mien1, Otub1, Vdac1 and Vegfa
   genes. Upregulated genes have a log fold change over two and a maximum
   FPKM above 100. (B) VDAC1-related molecules and gene expression. TSPO,
   VDAC1 and ANT in the mitochondrial membrane form a complex (left panel)
   and control the intracellular level of reactive oxygen species.
   Expression levels of upregulated Tspo and Ant genes (right panel). (C)
   Western blot analysis for TSPO was performed using PC12 cells collected
   at 6, 12 and 24 h post-^211At-MABG treatment. TSPO was detected at
   18-19 kDa, and its expression was increased in a time-dependent manner.
   In A and B, error bars represent the standard deviation among the three
   replicates, and the symbols indicate the median value.

   Mien1 encodes migration and invasion enhancer 1 (MIEN1), a factor
   regulating cytoskeletal-focal adhesion dynamics [124]^32. Ovarian tumor
   domain- containing ubiquitin aldehyde-binding protein 1 (OTUB1), which
   is encoded by Otub1, has a bilateral character, both inhibiting and
   promoting tumor growth [125]^33^, [126]^34. Mien1 and Otub1 genes may
   be important targets for better estimating prognosis of ^211At-MABG
   therapy. Vdac1 and Vegfa are linked to feasibility analysis for
   imaging: a mitochondrial 18-kDa translocator protein (TSPO), which
   forms a complex with 32-kDa voltage-dependent anion channel 1 (VDAC1)
   [127]^35, is a well-studied drug target for PET imaging [128]^25 and we
   measured the large increases in gene and protein expressions (Figure
   [129]6B-[130]6C). Vascular endothelial growth factor A (VEGFA) is also
   a reported target for PET imaging [131]^34. The features of these 4
   potential ^211At-MABG- specific molecular biomarkers may help doctors
   in diagnosis and in predicting prognosis with ^211At- MABG therapy.

Discussion

   In this work, we applied RNA-seq and showed contrasting gene expression
   patterns between ^211At- MABG treatment and γ-ray irradiation.
   Moreover, analysis of representative DEGs uncovered key gene expression
   associated with ^211At-MABG therapeutic effects and novel biomarkers
   for response to ^211At-MABG therapy (Figure [132]7A).

Figure 7.

   [133]Figure 7
   [134]Open in a new tab

   Potential biomarkers of ^211At-MABG-therapy and low fluence α-particle
   irradiation. (A) The present RNA-seq analysis suggests that ^211At-MABG
   therapeutic effects are associated with p53-p21 signaling and
   inhibition of ubiquitination by OTUB1. Furthermore, our analysis
   demonstrated novel targets for ^211At-MABG-therapy. OTUB1 and MIEN1 may
   be of importance for therapy and VDAC1 (TSPO) and VEGF-A for imaging.
   (B) Different irradiation modes of ^60Co γ-rays and α-particles emitted
   from ^211At-MABG. Photon radiation therapy uses high fluence-rate
   irradiation conducted in a short time and causes a biological response
   in a dose-dependent manner. In contrast, ^211At-MABG treatment is a
   mode of low fluence α-particle irradiation. The fluence rate was “one
   α-particle emission in one hour on average” for the 10% iso-survival
   experiment. Representative DEG analysis showed long-lasting p53-p21
   signaling post-^211At-MABG treatment.

Potential biomarkers for imaging in ^211At-MABG therapy

   CT or MRI and ^123I-MIBG are typically used to identify metastatic PCC
   [135]^1^, [136]^2, and PET with ^18F-fluorodopamine, ^18F-fluorodopa,
   ^18F-fluorodeoxyglucose and ^11C-hydroxyephedrine can be used as
   alternatives to ^123I-MIBG or ^131I-MIBG [137]^1^, [138]^2. These
   imaging procedures can be applied to diagnosis—e.g., localization of
   metastatic PCC—in ^211At-MABG therapy. However, there is no target to
   date for measuring the α-particle-induced tumor response, that is, a
   TAT-response biomarker. In this study, we found three potential genes,
   Vdac1, Vegfa and Mmp2, that may shed light on this issue.

   Vegfa specifically responds to MABG (Figure [139]6A) and is well known
   as a key imaging biomarker for tumor angiogenesis, i.e. tumor biomarker
   [140]^36^, [141]^37. We found that it was also potential response
   biomarker for ^211At-MABG therapy. HIF-1 [142]^38 and MYC [143]^39
   activate Vegfa gene transcription. Myc gene may be not associated with
   tumor angiogenesis of MABG- treated PC12 cells, because of
   under-regulation of Myc gene expression (Figure [144]5D). There are
   many inhibitors of VEGF signaling; for example, sorafenib inhibits
   RAS/RAF/MEk/ERK effects on VEGFA [145]^37. Thus, inhibitor use and
   monitoring of VEGFA may be important for assessing tumor angiogenesis
   after ^211At-MABG therapy.

   Overexpression of matrix metalloproteinases, e.g., MMP2 or MMP9, aids
   in prognosis for breast cancer patients [146]^40, and in our study, the
   Mmp2 gene was found to be representative at 12 h post-treatment (Table
   [147]1). Accordingly, evaluation of MMP2 response may be also important
   for prognosis after ^211At-MABG therapy.

   Here, we propose Tspo as a potential imaging biomarker because it is
   related to VDAC1 [148]^33; TSPO is a well-known drug target for PET
   [149]^25. TSPO, VDAC1 and inner membrane adenine nucleotide transporter
   (ANT) form a complex at the mitochondrial membrane (Figure [150]6B)
   that functions as a REDOX regulator of cell mitophagy, inducing various
   oxidative stress responses, including apoptosis [151]^35. Levels of
   Tspo, Vdac1 and Ant gene expression were significantly increased
   (Figure [152]6A-[153]6B), and that of Tspo was initiated in the early
   phase after ^211At-MABG treatment and maintained over three time points
   in a dose-dependent manner (Figure [154]6A-[155]6B). We also carried
   out western blot analysis, and levels of TSPO, detected at 18-19 kDa,
   increased over time in the 10% survival experiment (Figure [156]6C).
   ^18F-N-fluoroacetyl- N-(2,5-dimethoxybenzyl)-2-phenoxyaniline was
   reported to be a candidate probe for quantitative assessment of TSPO
   expression [157]^25. As PET signals from metastatic PCC can likely be
   detected in vivo, our results suggest that the PET imaging approach of
   assessing TSPO expression may detect the ^211At- MABG-induced tumor
   response and can provide indirect evidence of ^211At-MABG localization.

Biomarkers for better prognosis after ^211At-MABG therapy

   Metastasis-related gene expression is important for prognosis. The
   present study showed metastasis- related genes, not only Adam8, Jam3,
   Mmp2 and Gdf15 listed in Table [158]1 but also Mien1 and Otub1, with
   high-level induction. ADAM8 promotes invasiveness and reduces patient
   survival in pancreatic cancer [159]^41. The peptide blocker BK-1361 may
   reduce the cellular activity of Adam8 [160]^42. In addition, inhibitors
   of MMP2 and GDF15 have been reported [161]^43^, [162]^44, and a
   therapeutic target of MIEN1, microRNA miR-136, has been proposed
   [163]^45. Inhibitory drug and therapeutic targets of JAM3 and OTUB1 are
   not still understood, and in particular, the therapeutic effects of
   OTUB1 remain controversial.

   OTUB1 is a deubiquitinating enzyme that stabilizes p53 by inhibiting
   the ubiquitin-conjugating enzyme (E2)-dependent ubiquitination involved
   in cell death and growth inhibition [164]^33. Despite a recent report
   that OTUB1 stabilizes MDMX and induces apoptosis [165]^46, OTUB1
   stabilizes proteins linked to cancer progression such as RAS and FOXM1
   via deubiquitination and promotes tumor growth and invasion
   [166]^34^,[167]^47. Indeed, high expression of OTUB1 is clinically
   associated with tumor invasion, metastasis, poor prognosis, and low
   patient survival [168]^34. These previous reports indicate that OTUB1
   has dual functions that promote or suppress cancer. As expression of
   OTUB1 was dramatically enhanced by ^211At-MABG treatment (Figure
   [169]6A), OTUB1 is possibly a key enzyme in the tumor response to
   ^211At-MABG therapy in PCC.

Radiobiological insights and therapeutic mechanism of ^211At-MABG therapy

   We should note that according to the half-life of 7.2 h, the cellular
   irradiation of α-particles lasted longer than 12 minutes for ^60Co
   γ-ray irradiation (Figure [170]7B). The “low fluence rate” of
   α-particles in ^211At-MABG treatment was approximately “one α-particle
   emission in one hour on average” for the 10% iso-survival experiment
   (Figure [171]S2D; Supplementary result). The low fluence rate may be
   linked to the avoidance of cell death because HR and non-homologous
   end-joining contributed to DSB repair in TAT [172]^48; our results also
   suggest the contribution of HR, Fam175a and Rnf169 (Table [173]1).
   These findings raise the question regarding how anti-tumor therapeutic
   effects are induced at a low fluence rate, and the answer may be the
   p53 gene. Despite the low α-particle fluence rate, the p53 signaling
   pathway was overexpressed based on enrichment analysis (Figure [174]4),
   and p53 target genes related to cell cycle checkpoints and cell death—
   Cdkn1a, Enc1, Tp53inp1, Ccng1, Plk2, Eda2r, Aen and Btg2—were ranked
   among the representative DEGs (Table [175]1-[176]2). In particular,
   gene expression for most of the p53-targeted genes—e.g., Cdkn1a in
   Figure [177]S7—persisted and gradually increased over 3 time points.
   Possible causes may be due to p53-Mdm2 feedback [178]^49 and OTUB1
   [179]^33^, [180]^47, both of which stabilize and sustain p53 signaling.
   Long-lasting p53-induced signaling may break the wall of the “low
   fluence rate” in PC12 tumor cell death to exert anti-tumor effects.

Study limitations

   In this work, we have used PC12 as a representative cell-line for
   targeted radionuclide therapy, because PC12 is an important model
   cell-line for malignant pheochromocytoma study [181]^15. There is also
   a long history of nuclear medicine studies. In fact, the sufficient
   results of PC12 have contributed to the preclinical study of MIBG
   therapy, e.g. the report of Rutgers et al. [182]^16. On the other hand,
   human pheochromocytoma cell lines recently established such as KAT45
   and hPhe01 may also be useful future target for this kind of study, as
   well as studies in vivo with Rat models to further validate the
   candidate response biomarkers.

   The present study was mainly based on RNA-seq analysis. Therefore,
   post-translational modifications (PTMs) such as glycosylation and
   phosphorylation were not examined. Further studies will be needed to
   address PTMs. In addition, we did not detect significant mutations in
   the p53 gene in PC12 cells, though mutations in p53 frequently occur in
   many cancers. The therapeutic molecular mechanism of p53-mutated tumor
   cells in TAT may be different from the present results, according to
   the type of mutation.

Conclusions

   Our analysis demonstrated that ^211At-MABG therapeutic effects are
   associated with a molecular mechanism that occurs via the p53-p21
   pathway, the Otub1 gene related to ubiquitin-mediated proteolysis and
   other representative highly ranked DEGs. Additionally, we found 4
   potential molecular biomarkers that can be used in molecular imaging
   and as therapeutic targets of ^211At-MABG-therapy.

Supplementary Material

   Supplementary figures and tables.
   [183]Click here for additional data file.^ (1.7MB, pdf)

Acknowledgments