Total RNA-sequencing Detects Transcriptomic Differences in Mouse Preosteoblast Response to Bulk Nanostructured Titanium.

Rational: Titanium is the most widely used alloy family in dental and orthopedic implants due to its natural ability to integrate into bone, but cytotoxic alloying elements are required in titanium for its mechanical properties to match the functionality of natural bone in high-load bearing applications. Recent advances in nanostructuring, such as Continuous Equal Channel Angular Pressing (C-ECAP), reduce the grain size and increase the strength of pure grades of titanium, thereby eliminating the need for cytotoxic elements and increasing cytocompatibilty as measured by traditional cell biology techniques. Transcriptomic profiling of cells grown on conventional coarse grain (CG) versus nanostructured ultrafine grain (UG) surfaces can simultaneously enhance our understanding of genomics and biomaterials, facilitating the development of a new generation of implantable materials. Objective: We initiated this pilot project to develop a workflow for transcriptional profiling of both protein coding and non-coding RNAs (ncRNAs) in cells grown on Ti discs that are characterized for grain size, 12.1 µm for CG Ti and 0.28 µm for UG Ti. Methods: Two Ti substrates were designed for cell growth in 24-well culture dishes, CG and UG. Mouse pre-osteoblasts were cultured on Ti disc for 72 hours. Cells were harvested and total-RNA extracted. Ribosomal RNA subtraction was performed prior to Illumina sequencing library construction. Single-end 50 base pairs reads were subjected to Illumina quality control and the resulting reads were mapped to the mouse genome and annotations with Genomic Short-read Nucleotide Alignment Program (GSNAP) and HTseq-count. Read count files were normalized and statistically analyzed with empirical analysis of digital gene expression data in R (EdgeR) and filtered for low expression levels using Zipf's Power Law. Transcripts with adequate coverage and significant expression differences were subjected to GO Biological Process term enrichment and network analyses with the Cytoscape App ClueGO. Results: Gene expression differences with an FDR=0.05 are limited to six genes, five of which are likely ncRNAs. Of the 432 transcripts with significant differences at p=0.05 117 form an interconnected network of genes and 12 GO biological processes (P(adj) = 0.05). Biological processes that connect to predominately upregulated genes include cell division, mRNA metabolic processes, chromatin modification and protein localization to the cytoskeleton, while downregulated processes include regulation of apoptosis, of p38MAK cascade, and response to molecules of bacterial origin. Conclusions: After 72 hours, the transcriptional profile of cells on UG Ti confirms more rapid cell division than on CG. These results demonstrate that informative transcriptional profiling of cells growing on substrates that differ only in their crystallographic structure can be accomplished by a workflow that includes standard cell biology techniques, transcriptomic profiling, data mining, and network analyses. Overall design: Sterile Ti discs, 13.5 mm in diameter, polished to a 2 nm average surface roughness (Ra), were arrayed in a non-tissue culture treated 24-well cell culture plate using a block design, with 12 CG and 12 UG disks. Wells were re-fed after 24 hours and after 72 hours Ti discs were transferred to new plates for standard cell harvesting protocols. The cells from four wells were consolidated into one sample resulting in three biological replicates for each sample. Consolidated samples were stored at -80°C in RNAlater and all subsequent steps were identical for all samples. Samples on CG were used as the control for UG-CG comparisons. This project was partially supported by grants from the National Center for Research Resources (5P20RR016480-12) and the National Institute of General Medical Sciences (8 P20 GM103451-12) from the National Institutes of Health.

立论依据：钛（Titanium）是牙科与骨科植入物中应用最广泛的合金体系，因其天然具备骨整合能力。但在高载荷承载场景下，为使钛合金力学性能匹配天然骨的功能需求，需添加具有细胞毒性的合金元素（cytotoxic alloying elements）。近年来，连续等通道转角挤压（Continuous Equal Channel Angular Pressing，C-ECAP）等纳米结构化技术可细化纯钛牌号的晶粒尺寸并提升其强度，从而无需再添加细胞毒性元素，同时通过传统细胞生物学技术检测可证实其细胞相容性（cytocompatibility）得到提升。对比常规粗晶粒（coarse grain，CG）与纳米结构化超细晶粒（ultrafine grain，UG）表面培养的细胞进行转录组分析（transcriptomic profiling），既能加深我们对基因组学（genomics）与生物材料（biomaterials）的理解，也可推动新一代可植入材料的开发。
研究目的：本预实验项目旨在建立一套工作流程，用于分析在钛圆盘上培养的细胞中编码蛋白RNA与非编码RNA（non-coding RNAs，ncRNAs）的转录组特征；所用钛圆盘的晶粒尺寸经表征：粗晶粒钛（CG Ti）为12.1 µm，超细晶粒钛（UG Ti）为0.28 µm。
实验方法：设计两种钛基底用于24孔培养板中的细胞培养，分别为CG组与UG组。将小鼠前成骨细胞（Mouse pre-osteoblasts）接种于钛圆盘上培养72小时，收集细胞并提取总RNA。在构建Illumina测序文库前，先进行核糖体RNA（Ribosomal RNA）去除。对单端50bp读段（Single-end 50 base pairs reads）进行Illumina质控，随后使用基因组短读段核苷酸比对程序（Genomic Short-read Nucleotide Alignment Program，GSNAP）与HTseq-count将读段比对至小鼠基因组并完成注释。对读段计数文件进行标准化处理，基于R语言的数字基因表达数据实证分析包（EdgeR）进行统计学分析，并通过齐夫幂律（Zipf's Power Law）过滤低表达转录本。选取覆盖度充足且表达差异显著的转录本，使用Cytoscape插件ClueGO进行基因本体（Gene Ontology，GO）生物学过程术语富集分析与网络分析。
实验结果：错误发现率（false discovery rate，FDR）=0.05时，仅6个基因存在表达差异，其中5个大概率为非编码RNA。在P=0.05水平下存在显著表达差异的432个转录本中，117个构成基因互作网络，且富集得到12个校正后P值=0.05的GO生物学过程。主要上调基因关联的生物学过程包括细胞分裂、mRNA代谢过程、染色质修饰以及蛋白质向细胞骨架的定位；而下调的生物学过程则包括细胞凋亡调控、p38MAK级联反应调控以及对细菌来源分子的应答。
研究结论：培养72小时后，UG组钛圆盘上培养的细胞转录组特征显示其细胞分裂速率较CG组更快。本研究结果证实，仅通过标准细胞生物学技术、转录组分析、数据挖掘与网络分析组成的工作流程，即可实现对仅晶体结构存在差异的基底上培养细胞的有信息价值的转录组分析。
整体实验设计：将直径13.5mm、抛光至平均表面粗糙度（Ra）为2nm的无菌钛圆盘，采用区组设计排布于未经细胞培养包被的24孔细胞培养板中，其中CG组与UG组圆盘各12个。分别于接种后24小时更换培养基，培养至72小时后将钛圆盘转移至新培养板，遵循标准细胞收集流程获取细胞。将4个孔的细胞合并为1个样本，每个样本设置3个生物学重复。合并后的样本置于RNAlater保存液中并于-80℃冷藏，所有样本的后续实验步骤均保持一致。以CG组样本作为UG组与CG组比较的对照。本项目部分受美国国立卫生研究院（National Institutes of Health，NIH）下属国家研究资源中心（编号5P20RR016480-12）与国家普通医学科学研究所（编号8 P20 GM103451-12）的资助。