Experimental data demonstrating the effects of silver nanoparticles on basement membrane gene and protein expression in cultured colon, mammary and bronchial epithelia

This data article is related to the research article entitled “Silver nanoparticles alter epithelial basement membrane integrity, cell adhesion molecule expression and TGF-beta secretion”, available in the journal Nanomedicine: Nanotechnology, Biology, and Medicine [1]. This Data in Brief consists of data that describe changes in the expression of basement membrane (BM)-associated genes and proteins in three non-transformed epithelial cell lines following acute (6 h) and chronic (24 h plus 7-day chase) exposure to silver nanoparticles (AgNPs). Human BEAS2B (lung), MCF10AI (breast), and CCD-18Co (colon) cultured epithelia were analyzed for protein expression by LC-MS/MS and for gene expression by pathway-focused QRT-PCR arrays of 168 focal adhesion, integrin, and extracellular matrix (ECM) genes known to be localized to the plasma membrane, the BM/ECM, or secreted into the extracellular space. Ingenuity pathway analysis (IPA) of combined gene and protein expression datasets was then used to predict canonical pathways affected by AgNP exposure.


Value of the data
The majority of data describing in vitro silver nanoparticle (AgNP) toxicity has been generated from transformed (cancer) cell lines. In contrast, the data provided in this article characterizes AgNP-induced changes in the protein and gene expression of non-transformed ("normal") epithelial cell lines derived from lung, colon and breast tissue. Thus, these data provide baselines for comparison in future experiments that explore the pathophysiology of diseases induced by AgNP toxicity. More broadly, the control/baseline datasets of gene and protein expression from normal lung, colon and breast cell lines available in this article can be used for comparison a) with gene and protein expression in normal or diseased lung, colon and breast tissue in vivo, and b) with the same or similar cell lines tested in vitro and in vivo that have been altered by disease, genetic engineering or exposure to pathogens, therapeutic agents or toxins. The datasets available in this article identify panels of genes and proteins affected by AgNPs which embed in the extracellular matrix (ECM) and basement membrane (BM). These data have the potential to inform experiments to determine the outside-in/extracellular signaling effects of AgNP exposure on epithelial cells that support AgNP-stimulated intracellular signaling events and overall organ function. Most published datasets on molecular changes in the ECM include either proteomic or transcriptomic data, but not both, from a variety of species, tissue sources and/or methods of sample preparation. Data provided here include the expression of both proteins and genes associated with the ECM. Comparison of protein and gene expression data can be used to reveal post-transcriptional processes such as those that regulate ECM composition during tissue repair in murine lung [2]. These in vitro data provide a starting point for future, more clinically relevant in vivo experiments that can address the complex interactions between the ECM and diverse cell types in surrounding tissues [c.f. Ref [3]]. and 8 days in the three epithelial cell lines based on data in Tables 1 and 2. Table 4 provides QPCR array data of acute and chronic AgNP-induced changes in basement membrane-associated genes in BEAS2B (lung), MCF10AI (breast), and CCD-18Co (colon) cultured epithelia. Tables 5e7 summarize IPA analysis derived from LC-MS/MS protein data of AgNP-treated colon, lung and mammary epithelia, respectively. Figs. 1 and 2 illustrate the top canonical pathways and primary causal networks, respectively, associated with acute AgNP exposure in the three epithelial cell lines based on LC-MS/MS analysis data. Figs. 3 and 4 display top canonical pathways and primary causal networks, respectively, associated with chronic AgNP exposure in the three epithelial cell lines based on LC-MS/MS analysis data. Figs. 5 and 6 display top canonical pathways and primary causal networks, respectively, associated with changes between acute and chronic AgNP exposure in the three epithelial cell lines based on LC-MS/MS analysis data. Fig. 7AeF display top canonical pathways based on QPCR pathway-directed microarray data from the three epithelial cell lines exposed to acute (A-C) and chronic (D-F) AgNP; Fig. 7G-L display primary causal networks based on QPCR pathway-directed microarray data from the three epithelial cell lines exposed to acute (G-I) and chronic (J-L) AgNP. Supplementary data includes Table S1 (consisting of raw MS LFQ intensity values used to generate Tables 1e3), and PDF files of IPA summaries of LC-MS/MS data from all three cell lines at all three AgNP exposure conditions (acute, chronic and changes over time (OT)), which were used to generate Tables 5e7 2. Experimental design, materials and methods

Cell Culture
MCF10A, BEAS2B, and CCD-18Co cell models were obtained from American Type Culture Collection (Manassas, VA). Cells were cultured as directed by the supplier and all epithelial cells were cultured in "standard growth medium" (DMEM containing 10% FBS) during AgNP exposure for consistency.

AgNP Exposure
Each cell line was plated at concentrations predetermined to establish a confluent cell monolayer within three days. Cells were then cell cycle synchronized overnight in serum-starved medium, and then treated with 40 nm AgNP or AgNP diluent in standard growth medium for 6 or 24 hr. For chronic AgNP treatment, serum-starved confluent cells were treated for 24 hr with 40 nm AgNP or diluent (control cells) and then replaced with fresh medium (-AgNP) and cultured for eight days, with changes in medium every 2e3 days as needed.

Filter-aided Protein Sample Preparation
Cell pellets were subjected to filter-aided sample preparation [4] for protein cleanup using Vivacon 30,000 kDa molecular weight cutoff filters and reconstituted with 100 mL of 50 mM ammonium bicarbonate solution. Tryptic digestion was carried out by hydrating lyophilized trypsin to a stock solution of 1.0 mg/mL with 0.01% acetic acid in water; trypsin was added to the sample mixture at a 1:50 (v/ v) ratio and then incubated at 37 C for 4 hr. After digestion, peptides were acidified with HCl to a final concentration of 250 mM (pH 3) and aliquoted for LC-MS.

Nanoflow Liquid Chromatography (LC)
Pico-frit columns were purchased from New Objective (Woburn, MA) and packed to a length of 20e30 cm with reverse phase ReproSil-Pur 120 C-18-AQ 3.0 mm particles (Dr. Maisch GmbH HPLC, Ammerbuch-Entringen, Germany). Peptide separation was achieved on the column by injecting 2.0 mL of sample and using a gradient of mobile phase A (98.0% water, 2.0% acetonitrile, and 0.1% formic acid) and mobile phase B (80.0% acetonitrile, 20.0% water 0.1% formic acid). The LC method consists of a 120       min gradient with a linear ramp from 0.0% B to 40.0% B, a 1 min ramp to 100% B which is held for 6 min (123e129 minutes), followed by equilibration of the column at 0.0% B (130e140 minutes) running at a constant flow rate of 300 nL/min.

Orbitrap Mass Spectrometry (MS)
Orbitrap tandem mass spectrometry was performed using a Thermo Scientific Q-Exactive HF (Bremen, Germany) in a top 20 data dependent acquisition mode (DDA), where the 20 most abundant precursors were selected for fragmentation per full scan. MS1 and MS2 scans were performed at a resolving power of 120,000 and 15,000 at m/z 200, respectively. A dynamic exclusion window of 20 seconds was used to avoid repeated interrogation of abundant species. Automatic gain control was 1e6 and 1e5 for MS1 and MS2 scans, respectively. Samples were run in random order, and a quality control BSA digest was run and monitored every fifth injection to ensure proper LC-MS/MS reproducibility using AutoQC [5].

Protein Identification
Resulting raw data was loaded into MaxQuant (Version 1.5.6.0) [6], wherein MS/MS spectra were searched against a human proteome FASTA file downloaded from the Swiss-Prot protein database. The search included variable modifications of methionine oxidation and N-terminal acetylation, and fixed modification of cysteine carbamidomethylation. Peptides of a minimum of eight amino-acids and a maximum of two missed cleavages were allowed for the analysis. The peptide and protein identification false discovery rate (FDR) was set to 0.01. The resulting proteinGroups.txt data was imported into Perseus (Version 1.6.2.1) [7]. Here, reverse proteins or those only identified by site were filtered out. Next, the LFQ Intensity data was log 2 transformed and those proteins that did not produce a valid value in a minimum of two out of the three replicates in at least one group (i.e., proteins that were only detected once within a triplicate group), were filtered out. All remaining missing values were imputed using a normal distribution with a width of 0.3 and downshifted 1.8 standard deviations. Contaminants were then removed and group comparisons were performed using a two-sample student's t-test utilizing a Benjamini-Hochberg FDR calculation set to 0.05 for truncation. Tables 1 and 2 summarize changes in protein expression induced by acute and chronic AgNP treatment, respectively, of BEAS2B (lung), MCF10AI (breast), and CCD-18Co (colon) cultured epithelia as determined by LC-MS/MS. Table 3 summarizes changes in protein expression between 6 hours and 8 days in the three epithelial cell lines.

Quantitative Real Time PCR Arrays
Total RNA was isolated using the RNeasy kit (Qiagen; Hilden, Germany). RNA was reverse transcribed using the RT 2 First Strand Kit, and RT 2 Profiler PCR arrays (for focal adhesions, ECM and adhesion molecules) were used according to the manufacturer's instructions. Relative fold changes in gene expression were determined via the DD CT method using online analysis tools provided by Qiagen. Genes altered by acute and chronic AgNP exposure in BEAS2B (lung), MCF10AI (breast), and CCD-18Co (colon) cultured epithelia are found in Table 4.

Ingenuity pathway analysis (IPA)
Protein and mRNA datasets were imported into IPA for functional analysis (Qiagen, https://www. qiagenbioinformatics.com/products/ingenuity-pathway-analysis). The most significant networks and canonical pathways were predicted in IPA using restrictive statistical parameters to identify     1. Functional analysis of proteins altered by acute AgNP exposure. Colon, lung, and mammary cell models (CCD-18Co, BEAS2B, and MCF10AI, respectively) were exposed to AgNP for 6 h and changes in protein abundance was detected by LC MS/MS. Top canonical pathways were identified and generated via IPA. Histograms show the top significant canonical pathways with each respective elog (p-value) on horizontal axis during AgNP exposure. Threshold z-score for p-value is indicated with a horizontal orange line. Fig. 2. Functional analysis of proteins altered by acute AgNP exposure. Colon, lung, and mammary cell models (CCD-18Co, BEAS2B, and MCF10AI, respectively) were exposed to AgNP for 6 h and changes in protein abundance were detected by LC MS/MS. Primary causal signaling networks engaged during AgNP exposure as predicted by IPA. Lines and arrows between nodes represent direct (solid lines) and indirect (dashed lines) interactions between proteins. Red and green indicate up or down-regulation, respectively, and intensity of color indicates degree of regulation. Fig. 3. Functional analysis of proteins altered by chronic AgNP exposure. Colon, lung, and mammary cell models (CCD-18Co, BEAS2B, and MCF10AI, respectively) were exposed to AgNP for 8 days and changes in protein abundance was detected by LC MS/MS. Top canonical pathways were identified and generated via IPA. Histograms show the top significant canonical pathways with each respective elog (p-value) on horizontal axis during AgNP exposure. Threshold z-score for p-value is indicated with a horizontal orange line. Fig. 4. Functional analysis of proteins altered by chronic AgNP exposure. Colon, lung, and mammary cell models (CCD-18Co, BEAS2B, and MCF10AI, respectively) were exposed to AgNP for 8 days and changes in protein abundance were detected by LC MS/MS. Primary causal signaling networks engaged during AgNP exposure as predicted by IPA. Lines and arrows between nodes represent direct (solid lines) and indirect (dashed lines) interactions between proteins. Red and green indicate up or down-regulation, respectively, and intensity of color indicates degree of regulation. , and MCF10AI, respectively) were exposed to AgNP for 6 h and 8 days, and changes in protein abundance over time were detected by LC MS/MS. Top canonical pathways were identified and generated via IPA. Histograms show the top significant canonical pathways with each respective elog (p-value) on horizontal axis during AgNP exposure. Threshold z-score for p-value is indicated with a horizontal orange line. Fig. 6. Functional analysis of proteins altered by AgNP exposure over time. Colon, lung, and mammary cell models (CCD-18Co, BEAS2B, and MCF10AI, respectively) were exposed to AgNP for 6 h and 8 days and changes in protein abundance was detected by LC MS/MS. Primary causal signaling networks engaged during AgNP exposure as predicted by IPA. Lines and arrows between nodes represent direct (solid lines) and indirect (dashed lines) interactions between proteins. Red and green indicate up or downregulation, respectively, and intensity of color indicates degree of regulation.
pathways affected by significantly altered proteins or mRNAs. Algorithms defining networks and pathways are drawn from the Ingenuity Knowledge Base, a large, manually curated collection of nearly 5 million findings from the biomedical literature or integrated from third-party databases [8].
Canonical pathways classify molecules in the given dataset as per their reported ultimate biological function. Pathway significance is indicated by the number of molecules represented in the provided dataset with respect to the total number of identified molecules reported to affect the specific biological function. In the representative figures, calculated z-scores indicate top canonical pathways based on altered protein levels for the three epithelial cell lines exposed to acute (Fig. 1) and chronic (Fig. 3) AgNP exposure, as well as analysis of changes in protein expression between acute and chronic levels (Fig. 5). The ratio (orange dots connected by a line) indicates the ratio of proteins from the dataset that map to the pathway divided by the total number of genes that map to the same pathway. Primary causal network analysis of acute (Fig. 2) and chronic (Fig. 4) AgNP exposure and analysis of changes in protein expression between acute and chronic levels (Fig. 6) draws from approximately 40,000 nodes that represent mammalian genes and their products, chemical compounds, microRNA molecules and biological functions. Nodes are connected by approximately 1,480,000 edges representing experimentally observed causeeeffect relationships that relate to expression, transcription, activation, molecular modification and transport as well as binding events. Top canonical pathways and primary causal networks were also determined from QPCR pathway directed microarray data (Fig. 7). Fig. 7. Analysis of extracellular matrix, integrin, and focal adhesion genes altered by AgNP exposure. Colon, lung, and mammary cell models (CCD-18Co, BEAS2B, and MCF10AI, respectively) were exposed to AgNP and changes in gene expression identified by pathway-focused QPCR arrays. Functional analyses of AgNP-induced changes were generated via IPA. Histograms show the top significant canonical pathways with each respective elog (p-value) on horizontal axis during acute (AeC) and chronic (DeF) AgNP exposure. Threshold z-score for p-value is indicated with a horizontal orange line. (C, D) Primary causal signaling networks engaged during acute (GeI) and chronic (JeL) AgNP exposure as predicted by IPA. Lines and arrows between nodes represent direct (solid lines) and indirect (dashed lines) interactions between proteins. Red and green indicate up or down-regulation, respectively, and intensity of color indicates degree of regulation.