Spatial-Resolution Cell Type Proteome Profiling of Cancer Tissue by

Subscriber access provided by UNIV OF DURHAM

Spatial-resolution Cell Type Proteome Profiling of Cancer Tissue by Fully Integrated Proteomics Technology Ruilian Xu, Jun Tang, Quantong Deng, Wan He, Xiujie Sun, Ligang Xia, Zhiqiang Cheng, Lisheng He, Shuyuan You, Jintao Hu, Yuxiang Fu, Jian Zhu, Yixin Chen, Weina Gao, An He, Zhengyu Guo, Lin Lin, Hua Li, Chaofeng Hu, and Ruijun Tian Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.8b00596 • Publication Date (Web): 11 Apr 2018 Downloaded from http://pubs.acs.org on April 11, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Basic medicine, Tian, Ruijun; South University of Science and Technology of China, Department of Chemistry; Guangdong Provincial Key Laboratory of Cell Microenvironment and Disease Research

ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Spatial-resolution Cell Type Proteome Profiling of Cancer Tissue by Fully Integrated Proteomics Technology Ruilian Xu†,§, Jun Tang†,§,¶,ƒ, Quantong Deng§, Wan He§, Xiujie Sun‡, Ligang Xia§, Zhiqiang Cheng§, Lisheng He§, Shuyuan You§, Jintao Hu§, Yuxiang Fu§, Jian Zhu§, Yixin Chen§, Weina Gao‡, An He‡, Zhengyu Guo‡, Lin Lin‡, Hua Li‡, Chaofeng Hu¶, Ruijun Tian*,‡,ǂ §Shenzhen People’s Hospital, The Second Clinical Medical College of Jinan University, Shenzhen 518020and Department of Chemistry, South University of Science and Technology, Shenzhen 518055, China; ‡Department of Chemistry, South University of Science and Technology, Shenzhen 518055, China ¶Department of Pathophysiology, School of Basic medicine, Jinan University, Guangzhou 510632, China ƒIntegrated Chinese and Western Medicine Postdoctoral Research Station, Jinan University, Guangzhou 510632, China ǂGuangdong Provincial Key Laboratory of Cell Microenvironment and Disease Research, Shenzhen 518055, China *

E-mail: [email protected]. Phone: +86-755-88018905.


Page 2 of 27

Page 3 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


ABSTRACT Increasing attention has been focused on cell type proteome profiling for understanding the heterogeneous multicellular microenvironment in tissue samples. However, current cell type proteome profiling methods need large amount of starting materials which preclude their application to clinical tumor specimens with limited access. Here, by seamlessly combining laser capture microdissection and integrated proteomics sample preparation technology SISPROT, specific cell types in tumor samples could be precisely dissected with single cell resolution and processed for high-sensitive proteome profiling. Sample loss and contamination due to the multiple transfer steps are significantly reduced by the full integration and non-contacting design. H&E staining dyes which are necessary for cell type investigation could be selectively removed by the unique two-stage design of the spintip device. This easy-to-use proteome profiling technology achieved high sensitivity with the identification of more than 500 proteins from only 0.1 mm2 and 10 µm thickness colon cancer tissue section. The first cell type proteome profiling of four cell types from one colon tumor and surrounding normal tissue, including cancer cells, enterocytes, lymphocytes and smooth muscle cells was obtained. 5271, 4691, 4876 and 2140 protein groups were identified, respectively, from tissue section of only 5 mm2 and 10 µm thickness. Furthermore, spatially resolved proteome distribution profiles of enterocytes, lymphocytes and smooth muscle cells on the same tissue slices and across four consecutive sections with micrometer distance were successfully achieved. This fully integrated proteomics technology, termed LCM-SISPROT, is therefore promising for spatial-resolution cell type proteome profiling of tumor microenvironment with minute amount of clinical starting materials.



INTRODUCTION Organs and tissues are the functional units of human body and are one of the major sample sources as compared with cells and body fluids for biomedical studies. Mass spectrometry (MS)based proteomic studies typically performed on whole tissues have produced rich knowledge for better understanding the complex multicellular biological systems.1-3 As tissues often comprise wide range of cell types located at disparate anatomical regions with significant heterogeneity, such type of proteomic analysis inevitably produces an averaging effect along with the loss of spatial information that hinder the elucidation of deeper biological insights.4 To avoid the tissue heterogeneity and better study the local cell microenvironment, it is therefore essential and urgently needed for global proteome profiling of specific cell types in spatially refined tissue regions. Typically, well-established cancer cell lines are used for large-scale cell-type proteome profiling5,6. However, proteomics data obtained from these cell lines cannot fully represent and dissect the protein machinery happened in complex organ or tissue. Recently, spatial resolution primary cell type proteome profiling has drawn great attention in the proteomics society, which could be generally achieved by two types of approaches. One of the approach is based on isolating and culturing different cell types from organs and tissues. Azimifar et. al. investigated the biological function of mouse liver by isolating five cell types from mouse liver tissue and successfully providing deep proteome profiling of the individual hepatic cell types.7 Another investigation by Chen et al. adopted similar approach for studying the cell-type-resolved liver proteome and secretome of four hepatic cell types isolated from mouse liver tissue.8 However, these procedures inherently lose all spatial information, which is


Page 4 of 27

Page 5 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


critical for understanding the complex multicellular biological processes in local cell microenvironment. The other cell type proteome profiling approach is based on the dissection of different tissue region from relative large tissue or organs, such as brain9, heart10, and skeletal muscle.11 Sharma et al. performed brain proteome profiling of ten disparate anatomical regions by mouse brain tissue dissection and four isolated primary neurons including oligodendrocytes, astrocytes, microglia and cortical neurons.9 Doll et. al. established proteome maps of 16 different dissected regions of heart and three major primary cardiac cell types including cardiac fibroblasts, smooth muscle cells, and endothelial cells.10 However, all these studies require large amounts of starting materials which preclude the availability of these proteomic methods for clinical tissue specimens, especially for biopsy samples with very limited access. Patient tissue samples are one of the most commonly stored clinical resources for pathological investigation. Both formalin fixed paraffin embedded (FFPE) and fresh frozen tissues are commonly

kept

and

made

as

tissue

slices

for

hematoxylin-eosin

(H&E)

and

immunohistochemical staining. Due to significant chemical cross-linking and potential degradation during room temperature storage, fresh frozen tissues are preferred for downstream biological investigation when samples are available. For example, by combining with laser microdissection technology (LCM) with single cell resolution, refined fresh frozen tissue regions and cell types could be dissected out for further investigation such as by using genomic sequencing.12 Recently, MS-based proteomics has also been applied for studying the tissue section harvested by LCM with size down to 1 mm2 and 10 µm thickness.13 By adopting systematically optimized tissue extraction buffer and in-solution digestion, about 300 protein groups could be successfully identified. Although the proteome coverage of this technology is largely limited, the combination of MS-based proteomics and LCM technology is promising for



cell type proteome profiling directly from mass-limited clinical samples. Increasing attention has been focus on enhancing the sensitivity of LCM-based proteomics workflow by adopting integrated proteomics sample preparation technologies. For example, filter-aided sample preparation (FASP) developed by Mann group has become a popular method for general proteomics sample preparation and has been successfully used for proteome profiling of FFPE colon cancer tissue with relative large section size of 175 nL volume(~17 mm2 and 10 µm thickness.14,15 Another integrated proteome preparation method, termed single-pot solidphase-enhanced sample preparation (SP3), were also used for deep proteome profiling of FFPE ovarian tissues with large section size of ~1cm2 and 10 µm thickness.16,17 Although both FASP and SP3 technique have been successfully applied for LCM-harvested clinical tissue samples, especially with FFPE treatment, their performance for very limited tissue section is largely unknown. More importantly, H&E staining dyes for visualizing specific tissue regions is necessary for LCM dissection but will inevitably result in protein sample and MS contamination. Unfortunately, dye removal has not been fully considered in these integrated proteomics workflows yet. Recently, we developed the simple and integrated spintip-based proteomics technology (termed SISPROT) that enabled seamless integration of multiple steps of proteomics sample preparation, desalting and high-pH reversed phase (RP) fractionation into a single spintip device.18 The full integration design of the SISPROT technology has been approved to significantly increase the proteome profiling sensitivity and throughput for different sample types, such as stem cells18, gut microbiota19 and human plasma20. In this study, we aimed to further improve the proteome profiling sensitivity by adopting much smaller spintip device and systematically optimizing the integration of the LCM and SISPROT technology. Based on the unique SCX plus C18 packing


Page 6 of 27

Page 7 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


design of the SISPROT technology, we developed a new protocol for efficiently removing the H&E staining dyes while retain the protein contents for minimized SISPROT operation including high pH RP fractionation. The newly developed LCM-SISPROT technology was successfully applied for both cell type proteome and spatially resolved proteome profiling of three cell types dissection from colon cancer tissues for the first time.

EXPERIMENTAL SECTION Frozen Tissue Section Preparation and Laser Microdissection. Clinical colon specimens of cancerous tissue and surrounding normal tissue were obtained from the resected colon tissue after surgery, and the surrounding normal tissue was taken from the resection margin furthest away from the tumor region. All the specimens were frozen in liquid nitrogen immediately after collection and kept frozen in liquid nitrogen for further use. Fresh frozen tissues were embedded by Tissue OCT-Freeze Medium (Sakura Finetek USA, Inc.) in -20 oC and sliced into sections with thickness of 5, 10 and 15 µm by using Leica CM 1900 cryostat (Leica) at −20 oC. The obtained sections were then flat mounted onto a membrane-coated glass slides (2.0 µm, PENmembrane, Leica) following with frozen methanol fixing at -20 oC, hematoxylin-eosin staining (Baso), and dehydration via an ethanol series and air dryness. Tissue sections were scanned using Leica DM 2500 microscope at 5×magnification prior to LCM dissection. The obtained tissue slides were evaluated by pathological and cell type examination firstly and subjected to cell type dissection using a Leica LMD system (Leica LMD7000). The defined tissue regions were firstly marked under a 5× objective microscope and dissected the periphery of the selected section. The entire selected sections were then propelled onto the cap of 0.2 mL



EP tube (Axygen) by using a final laser pulse. LCM tissue slices with corresponding area ranging from 0.05 mm2 to 5 mm2 and with the thickness of 5 µm, 10 µm or 20 µm were collected for method optimization. LCM tissue slices from both cancerous and surrounding normal tissues with the area of 5 mm2 and 10 µm thickness were prepared for cell type proteome profiling experiment. Four different cell types (i.e. enterocytes, lymphocytes, smooth muscle cells and cancer cells, which generally derived from epithelial cells21) with well-defined morphology characteristics were carefully selected and dissected with the supervision of experienced pathologist for colon cancer. The percentage of cell type number in the four cell type regions under a 40× objective microscope were obtained by counting the number of corresponding cell type manually under the supervision of experienced pathologist for colon cancer as well. For spatially resolved cell type proteome profiling experiment, enterocytes, smooth muscle cells and lymphocytes were dissected from the same tissue section, respectively. Four consecutive slices from normal tissue were sliced with the thickness of 10 µm. Proteomics Sample Preparation. Dissected tissue sections in the cap of 0.2 mL EP tube were directly transferred into the same tube and lysed in 20 µL buffer containing 10 mM HEPES, pH 7.4, 150 mM NaCl, 600 mM guanidine HCl, 1% (w/v) DDM, 0.5% (w/v) PEG 2W and protease inhibitor mixture (1 mM EDTA, 1 mM PMSF, 1 µg/mL leupeptin, 1 µg/mL pepstatin, and 1 µg/mL aprotinin) assisted with non-contacting sonication. Protein concentration was determined by Pierce Micro BCA Kit (Thermo).22 The obtained tissue lysate was processed by using optimized SISPROT protocol with improvement for H&E staining dyes removal.18 Briefly, the samples were firstly acidified to pH 2−3 and loaded onto 200 µL or 10µLspintip device packed with one plug of C18 disk (3 M Empore, U.S.A.) and 0.6 mg of 20 µm POROS SCX beads (Applied Biosystems, U.S.A.) in tandem. H&E staining dyes, detergents and salt contaminants


Page 8 of 27

Page 9 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


were washed off by 20% (v/v) acetonitrile(ACN) in 8 mM potassium citrate buffer (pH 3) and an additional pure ACN wash. The remaining proteins were reduced by TCEP, alkylated by IAA and digested by trypsin (TPCK-treated, Sigma-Aldrich). The peptides were then transferred from the SCX beads to C18 disk with 200 mM ammonium formate (pH 10). For tissue section samples of 5 mm2, five high pH RP fractions with ACN concentration of 3%, 6%, 9%, 15%, 80% were obtained to reduce the sample complexity. Detailed protocol for LCM-SISPROT was shown in the supporting information. For comparison, in-solution digestion of the tissue slices was also performed according the reported protocol with minor modification.23 Briefly, detergent was removed from the lysates by acetone precipitation, and the proteins were digested with trypsin (Sigma-Aldrich) using insolution digest protocol (details in supporting information). The digested peptides were desalted using homemade C18StageTip.24 Nano-LC-MS/MS Analysis. The obtained samples were resuspended in 0.1% (v/v) formic acid (FA) and analyzed by an Orbitrap Fusion mass spectrometer coupled with an Easy-nLC 1000 (ThermoFisher Scientific). The LC separation was performed with an integrated spraytip column (100 µmi.d. × 20 cm) packed with 1.9 µm/120 Å ReproSil-Pur C18 resins (Dr. Maisch GmbH, Germany). The gradient solvent system consisted of solvent A [0.1% (v/v) FA in water] and solvent B [0.1% (v/v) FA in ACN]. 80% (v/v) of the peptide samples were loaded and separated at a flow rate of 250 nL/min. For 1 h gradient time, the solvent B was changed linearly as follows: 0 min, 3%; 2 min, 7%; 52 min, 22%; 62 min, 35%; 64 min, 90%; 70 min, 90%; 72 min, 3%; 80 min, 3%. For 2 h gradient, the solvent B as follows: 0 min, 3%; 2 min, 7%; 102 min, 22%; 122 min, 35%; 124 min, 90%; 130 min, 90%; 132 min, 3%;140 min, 3%. Full MS scans were performed in the Orbitrap mass analyzer over m/z range of 350−1550 with a mass



Page 10 of 27

resolution of 120000. The MS/MS spectra were acquired in data-dependent mode with a 3 s Top Speed method. Tandem MS was performed in the ion trap mass analyzer using an isolation window of 1.6 Da by quadrupole mass analyzer and HCD fragmentation with normalized collision energy of 30. The dynamic exclusion time was set as 60 s. Data Analysis. Raw data were searched against the human Uniprot fasta database (70332 entries, downloaded on Sep 29, 2016) using MaxQuant (version1.5.5.1). A maximum of two missed cleavages was allowed. Cysteine carbamidomethylation was set as fixed modification, while methionine oxidation, asparagine, and glutamine deamidation were set as variable modifications. FDR was set to 0.01. Quantification was performed by the same software.22,25 All statistical and bioinformatics analyses were performed with the Perseus software (version 1.5.5.3).26 Proteins identified with ≥ 2 peptides and 2 valid values in at least one group were reserved for further analysis. Protein groups identified with 3 valid values in at least one each group was reserved for spatial proteomics analysis. Missing values were assigned with an artificial value sampled from a normal constant.27 Data is presented as median ± SD. For cumulative protein intensities analysis, the proteins identified were ranged from the highest to the lowest abundance, such as A, B, C, D···Z,

each

cumulative

(A+B)/(A+B+C+D+···Z),

protein

intensity

was

calculated

(A+B+C)/(A+B+C+D+···Z),

by

A/(A+B+C+D+···Z),

(A+B+C+D)/(A+B+C+D+···Z)·and

(A+B+C+D+···Z)/(A+B+C+D+···Z)28. For heat map presentation, z-scored LFQ intensities after unsupervised hierarchical clustering were shown. Proteins are divided into four clusters showing Gene ontology (GO) annotations of biological process (BP) and cellular component (CC) were extracted from Uniprot database (GO release date: December 9, 2015). Top categorical annotations enriched after a Fisher’s exact test (p value 0.05) were shown.


Page 11 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


RESULTS AND DISCUSSION Design of the LCM-SISPROT Approach. The key characteristics of clinical cancer tissue samples are their significant heterogeneity of various cell types in the local tumor microenvironment and very limited access, especially for biopsy samples.29,30 Due to large amount of tissue samples are often needed, these recently developed spatial resolution and primary cell type proteomic approaches for studying big organs such as mouse liver, brain and heart are hard to be applied to limited amount of clinical tissue samples.7-10 Here, we designed an LCM-SISPROT approach for tackling this challenge. Due to the high heterogeneity of cancer tissue, the key challenge of the LCM-SISPROT approach is to specifically dissect single cell types without interference from neighboring cells. As shown in Fig. 1, we adopted single cell resolution tissue slicing and LCM technology for accurately dissecting specific cell types in the following two steps. We firstly made fresh frozen colon cancer tissue slide with single cell thickness by using tissue slicing platform with submicrometer resolution. We chose fresh frozen cancer tissues rather than FFPE cancer tissues because that the fresh frozen cancer tissues could preserve original protein machinery by avoiding significant chemical cross-linking and artificial degradation due to long term room temperature storage. More importantly, we could make fresh frozen tissue slide within 10 min by carefully optimizing the tissue slice fabrication procedure, while the fabrication of FFPE tissue slide often last for two or three days. We then stained the tissue slide with standard H&E dyes for morphological visualization and selected different cell types with the guidance of experienced oncologist of colon cancer. Each cell type region was dissected by LCM technology without disturbing the cells’ original morphology and contamination from neighboring cells. Noncontacting laser catapulting was used to transfer the dissected cell type regions into collection



tubes. Potential contamination is greatly avoided by the touchless design.31 The harvested tissue sections are subjected to proteomics sample preparation by using a dedicatedly designed minimized SISPROT spintip device with the following consideration. First, we developed and systematically optimized a minimized SISPROT method for significantly increasing the sensitivity of the overall workflow by avoiding sample loss due to transfer steps and nonspecific adsorption. The lysis process was designed to be operated in the original 0.2 mL tissue section collection tube assisted by non-contacting sonication rather than the traditional ultrasonic probe. According to a previous report, PEG polymer with an optimized concentration of 0.5% (w/v) was added to the lysis buffer for reducing nonspecific adsorption.32 As shown in Fig. 2A, we compared protein and peptide identification efficiency of the 200 µL and 10 µL volume spintip device for processing proteins lysed from tissue sections with the size of 1 mm2 and 0.1 mm2 (10 µm thickness), respectively. For 1 mm2 tissue sections, no obvious increase was observed in protein groups and peptides number between 200 µL and 10 µL spintip device. While for 0.1 mm2 tissue sections, a significant increase from 51 protein groups to 521 protein groups was achieved by changing 200 µL spintip device to 10 µL spintip device. These results confirmed that the minimized SISPROT method based on 10 µL spintip device is ideal for tissue sections with limited size largely because of greatly reduced nonspecific protein adsorption of limited amount of proteins. Second, we developed a unique function of the SISPROT spintip device for removing H&E staining dyes while keep the protein content intact. H&E staining of the tissue sections is typically required for morphological visualization and histological investigation.33 It is therefore unavoidable that the small molecular dyes (i.e. eosin and hematoxylin) will be introduced into the tissue lysate and ultimately contaminant the LC-MS system. Protein precipitation could


Page 12 of 27

Page 13 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


address the problem but is not favorable for minute amount of protein samples due to the significant sample loss. Due to their hydrophobic chemical structures (Fig. S1), the standard Stagetip C18 procedure could not separate the H&E staining dyes apart from the proteins. Here, we added additional dye removal step to the SISPROT method by taking advantages of the unique packing structure of SCX and C18 matrix in tandem. As shown in Fig. 1, the red H&E staining dyes retained on the SCX+C18 matrix are almost disappeared after pure ACN washing. SDS-PAGE analysis of the flow through fractions from the washing steps shows no distinguishable protein loss (Fig. S1C).This is largely because that the ionic interaction-based association between hematoxylin and eosin and proteins is largely broken due to the protonation of hematoxylin and eosin at pH 2-3 while proteins are mostly at positive charge state. When the sample is loaded onto the spintip device, the protein content could be easily trapped by the negatively charged SCX matrix, while the protonated hematoxylin and eosin with largely pure hydrophobic property will be washed off by pure ACN from both SCX and C18 matrix. LCM-SISPROT Has Superior Sensitivity. To test the performance of the LCM-SISPROT approach, we first compared it with in-solution digestion which was reported recently for processing limited amount of small fresh frozen uterine tissuesections.13 As shown in Fig. 2B, more than 1200 protein groups were identified from colon cancer tissue section of 1 mm2 and 10 µm thickness by our in-solution digestion method, while only ~300 proteins were identified by the reported in-solution digest method.13 In comparison, more than 2000 protein groups were identified from the same size of tissue section when the optimized SISPROT method was used. We reason that the integrated SISPROT method outperformed the in-solution digestion method not only because of its fully integrated sample preparation but also the dye removal function.



We further evaluated the sensitivity of the optimized LCM-SISPROT approach for processing tissue sections with different size and thickness. Tissue sections with the thickness of 5 µm, 10 µm and 20 µm were firstly evaluated while keeping the tissue area as1 mm2. Fig. 2C indicated that the identified protein groups and peptides increased dramatically when the tissue thickness increased from 5 µm to 10 µm and then reached to a plateau from 10 µm to 20 µm. Since the typical cell dimension is about 10 µm which is greatly helpful for isolating pure cell types, we therefore adopted 10 µm as an appropriate thickness for our cell type proteome profiling. Benefiting from the LCM technology with single cell resolution, we next processed tissue sections with different area while keeping the same 10 µm thickness. As shown in Fig. 2D, 149, 521, 1738 and 2061 protein groups were identified by single-shot proteomic analysis for LCM samples with the area of 0.05, 0.1, 0.5 and 1mm2, respectively, while 4458 protein groups were identified for LCM sample with the area of 5 mm2 by employing additional 5 step high pH RP fractionation on the same spintip device. It’s worth to mention that we also tried to extract proteins from LCM sample with the area of 10 mm2. However, no significant protein concentration difference was observed between 5 and 10 mm2 as measured by the micro BCA method (Fig. S2).34 This result is reasonable as increasing amount of non-protein components in larger tissue sections would influence the protein extraction efficiency of this dedicatedly design minimized proteomics workflow. Therefore, about 2.2 µg proteins extracted from 5 mm2 tissue section was adopted for global cell type proteome profiling. Cell type proteomics analysis. We exemplified the cell type proteome profiling technology by studying different cell types in colon cancer tissue. By careful morphological investigation of fresh frozen colon cancer tissue slices, four cell types, including enterocytes, lymphocytes, smooth muscle cells and cancer cells, were selected according to the following well-established


Page 14 of 27

Page 15 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


signatures: enterocytes are characterized as monolayer columnar cells; lymphocytes are characterized as dense arranged cells with basophilic cytoplasm that strongly stained with hematoxylin; smooth muscle cells are characterized as long shuttle cells; and adenocarcinoma is characterized by glandular formation.35 As shown in Fig. 3A, the percentages of positive cells are 65%, ˃90%,˃90% and 81% for enterocytes, lymphocytes, smooth muscle cells and cancer cells, respectively. The cell type region with accumulated total area of 5 mm2 (10 µm thickness) were dissected by using LCM technology, processed by the minimized SISPROT method, fractionated into 5 fractions by high-pH RP fractionation, analyzed on the LC-MS/MS system and the results were summarized in Tables S2. By combining the identified proteins from three tissue slices from the same patient sample, 5271, 4691, 4876 and 2140 protein groups were identified from cancer cells, enterocytes, lymphocytes and smooth muscle cells, respectively (Fig. 3B). Fig. 3C showed that the LCM-SISPROT approach has good label-free quantification precision with of Pearson correlation of above 0.9. In summary, we built up the first proteome databases for four important cell types for colon cancer with good proteome coverage and reproducibility. Interestingly, protein groups and peptides number obtained from smooth muscle cells is less than half of that obtained from three other cell types. This is largely caused by the highest dynamic range of smooth muscle cells36 (i.e. 5.6) as compared with three other cell types (Fig. 4A).37 We further analyzed the high abundant proteins of smooth muscle cells by ranking the identified proteins according to their cumulative protein intensities, which indicating the incremental proportion of a protein amount acounting for total protein amount (Fig. 4B).28 Four of the most abundant proteins including three smooth muscle cell specific proteins (i.e. DES, TAGLN and CNN1) account for ~50% of the total proteins of smooth muscle cells.38 The other five cell type specific marker proteins including CALD1, TAGLN2, CNN3, CNN2 and TAGLN3 were marked



red and found in the low abundance part of cumulative proteome profiling curve.38 In comparison, 10 to 23 of the most abundant proteins accounted only for about 25% of the total protein amount for cancer cells, enterocytes and lymphocytes (Fig. S3). These results were also confirmed by other recent reports.32 Beside for smooth muscle cells, 19, 1 and 2 marker proteins were also found in the global proteome profiling of cancer cells, enterocytes and lymphocytes (Fig. S3 and Table S1).32,39,40 We did further bioinformatic analysis for annotating the identified proteins from all four cell types. Principal-component analysis (PCA) on the proteome data of the four cells types shows clear separation indicating their significant signature difference (Fig. 4C). Consistent with their originality, lymphocytes were the most diverging cell type as one sample was obtained from colon cancer tissue (annotated as Lymphocyte T) while the other two samples were obtain from normal tissue (annotated as Lymphocyte N). The abundance distribution of all the identified proteins in the four cell types indicates the significant protein expression difference between different cell types while highly consistent expression between the triplicate analysis of the same cell types (Fig. 4D). We did further functional annotation by Gene Ontology analysis of both biological processes (GOBP) and cellular components (GOCC) for the identified proteins with distinct clustering profile in specific cell types. Extracellular matrix proteins were enriched in smooth muscle cells, which is consistent with the previous report that smooth muscle tissue includes extensive extracellular matrix.41 This is also indicated by the biological function annotation with enrichment for extracellular structure organization. The biological processes enriched for lymphocyte are mostly those involving regulation of immune process. For enterocytes, we found the enrichment for O-glycan processing which is related to the digestion and the absorption of nutrients.42 A range of membrane components were enriched in both


Page 16 of 27

Page 17 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


enterocyte and lymphocyte indicating the involvement of related membrane receptors in these highlighted biological processes. Lastly, identified proteins which are enriched in cancer cells are mainly involved in intracellular gene expression. Spatial resolution cell type proteome profiling. With the great heterogeneity of cancer tissue, it will be promising to analyze the spatial resolution cell type proteome in a vertical dimension. Mass spectrometry imaging is a well-established MS method for providing spatial distribution information of proteins in tissues. However, the proteome coverage of identified proteins is very limited and prone to matrix effects challenge for quantification.43,44 Here, we attempted to construct the cell type proteome in the vertical dimension of three different cell types located in the same tissue section and across four consecutive sections with distance of about 10 µm (Fig. 5A and Fig. S4). As summarized in Table S3, the protein groups and peptides numbers identified from enterocytes, lymphocytes and smooth muscle cells with the area of 5 mm2, 0.2 mm2 and 5 mm2 by a single shot MS analysis were obtained. The proteome coverage of ~3500 protein groups was obtained for enterocytes and lymphocytes in each tissue section within the variation range of 300 protein groups (Fig. 5B). While for smooth muscle cells, the proteome coverage of ~1000 protein groups was obtained due to the existence of high abundant proteins (Fig. 5B). Heatmap distribution of the commonly identified proteins over the four consecutive sections confirmed that most of the proteins has consistent expression level across micrometer spatial distance (Fig. 5C). However, there are significant number of proteins showed protein expression fluctuation across the spatial distance. As understanding of spatial molecular composition both in two dimensions and three dimensions is very important in the analysis of development and spread of tumor microenvironment, we hope the cell type proteome profiling in the vertical dimension will be useful in providing deeper insight into the complex biological mechanism of



cancerous processes.

CONCLUSIONS An easy-to-use and high-sensitive cell type proteome profiling method has been developed based on the tissue microdissection technology with single cell resolution and the integrated proteomics sample preparation technology SISPROT. Proteomic analysis sensitivity is greatly increased by the dedicatedly designed minimized SISPROT procedure. H&E staining dye contaminant was efficiently removed by the unique design of the SISPROT spintip device. The first proteome databases for four important colon tumor cell types, including cancer cells, enterocytes, lymphocytes and smooth muscle cells, were successfully established with only 5 mm2 and 10 µm thickness of starting materials. Spatial resolution cell type proteome profiling analysis across four consecutive tissue sections revealed distinguishable protein expression difference with micrometer spatial resolution. We expect that the LCM-SISPROT method become a unique and generic approach for spatial-resolution cell type proteome profiling of clinical tissue samples with limited access.

ASSOCIATED CONTENT Supporting Information The Supporting Information Available: [Experimental details, Figures S1−S4, Tables S-1 (PDF) and Tables S-2, S-3 (ZIP).]

AUTHOR INFORMATION


Page 18 of 27

Page 19 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Corresponding Author *E-mail: [email protected]. Phone: +86-755-88018905. Author Contributions †These authors contributed equally. Notes The authors declare no competing financial interest.

ACKNOWLEDGEMENTS We thank Dr. Lin Lin and Dr. Hua Li for their support of MS analysis; We thank Chief physician of gastrointestinal surgery Ligang Xia for the Specimens of colon cancer. We thank Chief physician Zhiqiang Cheng of pathology department, Pathology Technician Lisheng He and Clinical Pathologist Shuyuan You in Shenzhen People’s Hospital for their excellent technical assistance in making pathological sections and professional directions in interpreting pathological sections. This study was supported by grants from the foundation items: National Natural Science Foundation of China (No. 21575057), Guangdong Provincial Grants (2017B030301018 and 2016A030312016), Shenzhen Innovation of Science and Technology Commission

(No.

JCYJ20150901153557178,

JSGG20160301103415523

and

JCYJ20160229153100269), and Shenzhen Health and Family Planning Commission (SZXJ2017025).

REFERENCES



(1) Huttlin, E. L.; Jedrychowski, M. P.; Elias, J. E.; Goswami, T.; Rad, R.; Beausoleil, S. A.; Villen, J.; Haas, W.; Sowa, M. E.; Gygi, S. P. Cell2010, 143, 1174-1189. (2) Wilhelm, M.; Schlegl, J.; Hahne, H.; Gholami, A. M.; Lieberenz, M.; Savitski, M. M.; Ziegler, E.; Butzmann, L.; Gessulat, S.; Marx, H.; Mathieson, T.; Lemeer, S.; Schnatbaum, K.; Reimer, U.; Wenschuh, H.; Mollenhauer, M.; Slotta-Huspenina, J.; Boese, J. H.; Bantscheff, M.; Gerstmair, A., et al. Nature2014, 509, 582-587. (3) Kim, M. S.; Pinto, S. M.; Getnet, D.; Nirujogi, R. S.; Manda, S. S.; Chaerkady, R.; Madugundu, A. K.; Kelkar, D. S.; Isserlin, R.; Jain, S.; Thomas, J. K.; Muthusamy, B.; LealRojas, P.; Kumar, P.; Sahasrabuddhe, N. A.; Balakrishnan, L.; Advani, J.; George, B.; Renuse, S.; Selvan, L. D. N., et al. Nature2014, 509, 575-581. (4) Kehr, J. Curr. Opin. Plant Biol.2003, 6, 617-621. (5) Lawrence, R. T.; Perez, E. M.; Hernandez, D.; Miller, C. P.; Haas, K. M.; Irie, H. Y.; Lee, S. I.; Blau, C. A.; Villen, J. Cell Rep.2015, 11, 630-644. (6) Gholami, A. M.; Hahne, H.; Wu, Z.; Auer, F. J.; Meng, C.; Wilhelm, M.; Kuster, B. Cell Rep.2013, 4, 609-620. (7) Azimifar, S. B.; Nagaraj, N.; Cox, J.; Mann, M. Cell Metabolism2014, 20, 1076-1087. (8) Ding, C.; Li, Y. Y.; Guo, F. F.; Jiang, Y.; Ying, W. T.; Li, D.; Yang, D.; Xia, X.; Liu, W. L.; Zhao, Y.; He, Y. Z. G.; Li, X. Y.; Sun, W.; Liu, Q. M.; Song, L.; Zhen, B.; Zhang, P. M.; Qian, X. H.; Qin, J.; He, F. C. Mol. Cell. Proteomics2016, 15, 3190-3202. (9) Sharma, K.; Schmitt, S.; Bergner, C. G.; Tyanova, S.; Kannaiyan, N.; Manrique-Hoyos, N.; Kongi, K.; Cantuti, L.; Hanisch, U. K.; Philips, M. A.; Rossner, M. J.; Mann, M.; Simons, M. Nat. Neurosci.2015, 18, 1819-1831. (10) Doll, S.; Dressen, M.; Geyer, P. E.; Itzhak, D. N.; Braun, C.; Doppler, S. A. Nat. Commun.2017, 8, 1469-1481. (11) Murgia, M.; Toniolo, L.; Nagaraj, N.; Ciciliot, S.; Vindigni, V.; Schiaffino, S.; Reggiani, C.; Mann, M. Cell Rep. 2017, 19, 2396-2409. (12) Casasent, A. K.; Schalck, A.; Gao, R.; Sei, E.; Long, A.; Pangburn, W.; Casasent, T.; MericBernstam, F.; Edgerton, M. E.; Navin, N. E. Cell2018, 172, 205-217. (13) Dapic, I.; Uwugiaren, N.; Jansen, P. J.; Corthals, G. L. Anal. Chem. 2017, 89, 10769-10775. (14) Wisniewski, J. R.; Zougman, A.; Nagaraj, N.; Mann, M. Nat. Methods2009, 6, 359-362. (15) Wisniewski, J. R.; Ostasiewicz, P.; Dus, K.; Zielinska, D. F.; Gnad, F.; Mann, M. Mol. Syst.


Page 20 of 27

Page 21 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Biol.2012, 8, 611. (16) Hughes, C. S.; Foehr, S.; Garfield, D. A.; Furlong, E. E.; Steinmetz, L. M.; Krijgsveld, J. Mol. Syst. Biol.2014, 10, 1-10. (17) Hughes, C. S.; McConechy, M. K.; Cochrane, D. R.; Nazeran, T.; Karnezis, A. N.; Huntsman, D. G.; Morin, G. B. Sci. Rep.2016, 6, 34949. (18) Chen, W.; Wang, S.; Adhikari, S.; Deng, Z.; Wang, L.; Chen, L.; Ke, M.; Yang, P.; Tian, R. Anal. Chem.2016, 88, 4864-4871. (19) Zhang, X.; Chen, W.; Ning, Z.; Mayne, J.; Mack, D.; Stintzi, A.; Tian, R.; Figeys, D. Anal. Chem.2017, 89, 9407-9415. (20) Lin, L.; Zheng, J.; Yu, Q.; Chen, W.; Xing, J.; Chen, C.; Tian, R. J. Proteomics2018, 174, 916. (21) Wong, W. M.; Mandir, N.; Goodlad, R. A.; Wong, B. C. Y.; Garcia, S. B.; Lam, S. K.; Wright, N. A. Gut2002, 50, 212-217. (22) Cox, J.; Hein, M. Y.; Luber, C. A.; Paron, I.; Nagaraj, N.; Mann, M. Mol. Cell Proteomics.2014, 13, 2513-2526. (23) Borner, G. H.; Fielding, A. B. Cold Spring Harbor protocols2014, 1192-1195. (24) Naldrett, M. J.; Zeidler, R.; Wilson, K. E.; Kocourek, A. J. Biomol. Tech.2005, 16, 423-428. (25) Cox, J.; Mann, M. Nat. Biotechnol.2008, 26, 1367-1372. (26) Tyanova, S.; Temu, T.; Sinitcyn, P.; Carlson, A.; Hein, M. Y.; Geiger, T.; Mann, M.; Cox, J. Nat. Methods2016, 13, 731-740. (27) Lazar, C.; Gatto, L.; Ferro, M.; Bruley, C.; Burger, T. J. Proteome Res.2016, 15, 1116-1125. (28) Sharma, K.; Schmitt, S.; Bergner, C. G.; Tyanova, S.; Kannaiyan, N.; Manrique-Hoyos, N.; Kongi, K.; Cantuti, L.; Hanisch, U. K.; Philips, M. A.; Rossner, M. J.; Mann, M.; Simons, M. Nat. Neurosci.2015, 18, 1819-1831. (29) Quail, D. F.; Joyce, J. A. Nat. Med.2013, 19, 1423-1437. (30) Barcellos-Hoff, M. H.; Lyden, D.; Wang, T. C. Nat. Rev. Cancer2013, 13, 511-518. (31) Vandewoestyne, M.; Goossens, K.; Burvenich, C.; Van Soom, A.; Peelman, L.; Deforce, D. Anal. Biochem.2013, 439, 88-98. (32) Wisniewski, J. R.; Ostasiewicz, P.; Mann, M. J. Proteome Res.2011, 10, 3040-3049. (33) Chan, J. K. Int. J. Surg. Pathol.2014, 22, 12-32. (34) De Graaf, E. L.; Pellegrini, D.; McDonnell, L. A. J. Proteome Res.2016, 15, 4722-4730.



(35) Zou, Z. Z.; Li, J. C. Histology and Embrydogy, 8th ed.; People’s Medical Publishing House: Beijing, 2013. (36) Deshmukh, A. S.; Murgia, M.; Nagaraj, N.; Treebak, J. T.; Cox, J.; Mann, M. Mol. Cell. Proteomics2015, 14, 841-853. (37) Azimifar, S.B.; Nagaraj, N.; Cox, J.; Mann, M.Cell Metab.2014, 20, 1076-1087. (38) Robin, Y. M.; Penel, N.; Perot, G.; Neuville, A.; Velasco, V.; Ranchere-Vince, D.; Terrier, P.; Coindre, J. M. Mod. Pathol.2013, 26, 502-510. (39) Chougule, P.; Herlenius, G.; Hernandez, N. M.; Patil, P. B.; Xu, B.; Sumitran-Holgersson, S. Scand. J. Gastroenterol.2012, 47, 1334-1343. (40) Senovilla, L.; Vacchelli, E.; Galon, J.; Adjemian, S.; Eggermont, A.; Fridman, W. H.; Sautes-Fridman, C.; Ma, Y.; Tartour, E.; Zitvogel, L.; Kroemer, G.; Galluzzi, L. Oncoimmunology2012, 1, 1323-1343. (41) Eddinger, T. J. Anat. Rec.2014, 297, 1734-1746. (42) Parnis, S.; Nicoletti, C.; Ollendorff, V.; Massey-Harroche, D. J. Cell Physiol.2004, 198, 441-451. (43) Lanekoff, I.; Stevens, S. L.; Stenzel-Poore, M. P.; Laskin, J. Analyst2014, 139, 3528-3532. (44) Heeren, R. M.; Smith, D. F.; Stauber, J.; Kukrer-Kaletas, B.; MacAleese, L. J. Am. Soc. Mass Spectrom.2009, 20, 1006-1014.


Page 22 of 27

Page 23 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 1. Schematic workflow of the spatial-resolution cell type proteome profiling method for defined cell types in colon cancer: Can, cancer cells; Lym, lymphocytes; Ent, enterocytes; Mus, smooth muscle cells.



Figure 2. Optimization of the minimized SISPROT method. (A) Effect of spintip device volume; a, 200 µL; b, 10 µL. (B) Comparison between in-solution digestion and SISPROT-based integrated digestion; (C) Effect of tissue thickness for protein groups and peptides identification by single-shot 1 h LC-MS/MS analysis; 1 mm2 tissue area. (D) Effect of tissue area for protein groups and peptides identification by single-shot 1 h LC-MS/MS analysis or 5 high pH RP fractionation followed by 2 h LC-MS/MS analysis for 5 mm2tissue section; 10 µm tissue thickness.


Page 24 of 27

Page 25 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 3. Performance of the cell-type proteome profiling of colon cancer. (A) The morphology and cell type percentage of the four cell types in dissected tissue sections; 40 × objective microscope. (B) Identified protein groups in the triplicate analysis and combination of triplicate analysis of each cell types; 5 mm2 and 10 µm thickness of defined cell types. (C) Correlation of LFQ intensity of each identified proteins in the triplicate analysis.



Figure 4. Bioinformatic analysis of the four cell type proteomes. (A) The dynamic range of the four cell type proteomes. (B) Cumulative protein mass for the cell type of smooth muscle cells, the proteins identified were ranged from the highest to the lowest abundance. The total number of proteins constituting different quantiles were shown and the cell specific marker proteins were marked with gene names. (C) PCA distribution plot of the individual cell-type proteomes (n = 3 for each cell type); Lym-T, lymphocyte from tumor region; Lym-N, lymphocyte from normal tissue region. (D)Heatmap of protein groups differentially expressed across different cell types (n = 3 for each cell type). Top enriched GO categories for clusters are shown.


Page 26 of 27

Page 27 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 5. Spatial resolution cell type proteome profiling of enterocytes, lymphocytes and smooth muscle cells in a vertical dimension. (A) Four consecutive tissue sections with spatial distance of ~10 µm were prepared for dissecting the indicating cell types; (B) Identified protein groups and peptides from the three spatially resolved cell types; (C) Heatmaps of protein abundance distribution.



For TOC only


Page 28 of 27

Spatial-Resolution Cell Type Proteome Profiling of Cancer Tissue by

Recommend Documents