An array of 60,000 antibodies forproteome-scale antibody generation and target discovery
Zhaohui Wang, Yang Li (IMMUNOLOGY)
https://advances.sciencemag.org/content/6/11/eaax2271
摘要
抗體是闡明基因功能的關鍵。然而，目前還沒有價格低廉的蛋白質組規模抗體產生技術。為了解決這個問題，本文開發了蛋白質組表位標籤抗體庫（PETAL）及其陣列。PETAL由62,208個單克隆抗體（mAbs）組成，針對不同蛋白質組的15,199個肽。由於抗體的多特異性（一種固有的抗體特性），PETAL在自然界中含有大量蛋白質結合物。通過陣列篩選，發現了10000到20000個mAbs的特異性組合以靶向特定蛋白質組。在玉米和斑馬魚樣品中發現了表型特異的mAb蛋白對。從各自的蛋白組結合PETAL mAbs中鑑定了膜蛋白免疫螢光和流式細胞術mAbs和轉錄因子的染色質免疫沉澱測序mAbs。通過對腫瘤和正常組織細胞表面蛋白質組進行差異篩選，確定了抗體藥物偶聯物的內在化腫瘤抗原。通過與當前情況相比，花費很少時間和成本就可找到高親和力的mAbs，PETAL可以實現蛋白質組規模的抗體生成和靶標發現。
引言
在當前DNA測序技術不斷發展的能力推動下，已經對1300多種動物、496種植物以及許多其他物種的基因組進行了測序，代表了數百萬個基因，而且從G10K、i5k等項目來看，這一數字將更快增長【1】。為了了解這些基因的作用，需要探索基因編碼蛋白的功能，迫切需要抗體，特別是在蛋白質組水平上產生的可更新的單克隆抗體（mAbs）。人類蛋白質雜交瘤技術產生的mAbs一直被認為是發現診斷和治療靶點的最直接工具【2】。經典的治療靶點唾液酸化的Lewis Y是前列腺特異性膜抗原，最近，由細胞表面蛋白的mAbs發現了一個先前未知的多發性骨髓瘤靶點【3-5】。
儘管mAbs和基於mAb的發現非常重要，但是大規模mAbs的產生仍然困難重重，因為傳統的雜交瘤的發展非常耗時（從抗原開始4到6個月）、昂貴（每個抗原3000到8000美元）且難以擴展。此外，通過免疫產生mAbs通常需要一毫克純化抗原，這對許多蛋白質，特別是作為主要研究對象的膜蛋白來說是巨大挑戰。即使對於人類蛋白質，由於缺乏用於流式細胞術[螢光活化細胞分選（FACS）]和免疫螢光（IF）等應用的高質量抗體，6000種膜蛋白中的大多數尚未被直接用作診斷性或治療性靶點【6,7】。
人類蛋白質圖譜（HPA）為蛋白質組規模抗體的發展提供了一種選擇性新途徑。HPA已經產生了針對17000多個人類蛋白質的25000多個親和純化的多克隆抗體，其覆蓋了超過80%的人類蛋白質組【8,9】。然而，由於該項目需要大量的人力和資本資源，在需要蛋白質組規模抗體的大多數其他物種上複製HPA的成功不切實際。此外，多克隆HPA抗體不可再生，這些重新生成的抗體的質量難以保持一致。因此，對大多數序列基因組而言，蛋白質組規模抗體的產生一直難以捉摸。
在過去的幾十年裡，許多學者試圖通過改進傳統的雜交瘤方法，開發更好的體外重組抗體庫和更有效的篩選技術，來解決大規模抗體產生的高成本和低可擴展性問題【10-12】。在體外方法方面，20年前就有人嘗試繼續發展新的顯示技術和改進抗體庫設計和篩選方法【13-15】。然而，由於高成本的考慮，所建立的用於治療性抗體發現的合成抗體庫尚未用於大規模試劑產生。儘管有人試圖利用噬菌體展示庫產生研究性抗體，但將這些資源用於產生非臨床用途或非人類蛋白質組用途的抗體並不經濟【16】。
抗體微陣列是一個強大的平臺，可利用固定化抗體集合進行高通量、多路復用蛋白質分析【17,18】。利用抗體陣列技術，通過直接陣列篩選，可以實現低成本、快速的抗體發現。之前有方法使用由約10000個矽化物設計和合成的抗體片段組成的文庫來構建抗體陣列，用於重新發現抗體【19】。陣列庫能夠產生對治療性蛋白質靶點具有微極性結合親和力的先導抗體，這表明包含數萬個單獨抗體的空間定位庫應該足以發現抗體。然而，由於抗體親和力成熟方法和設計的需要額外這限制了這種合成抗體庫篩選法在常規研究親和力試劑開發和靶向分析中的作用，因此該方法未產生更廣泛影響。
本文提出了一個集成「工業規模」雜交瘤開發、抗體微陣列和親和蛋白質組學的系統，以克服以往蛋白質組規模抗體開發和靶標發現的挑戰。本文技術稱為蛋白質組表位標籤抗體庫（PETAL），其利用抗體的多特異性，這是抗體分子的一種固有特性，使其與大量具有高度親和力和特異性的與原始抗原無關的蛋白質結合。抗體多特異性可以以抗p24（HIV-1）肽mAb（CB4-1）作為範例【20,21】。其表位序列由5個與CB4-1結合的無關肽的關鍵相互作用殘基組成，在數百個異源蛋白質中被鑑定，並且經證明，那些可以獲得的蛋白質與CB4-1結合【22】。PETAL是一個小鼠mAb庫，由62208個mAbs組成，針對418個蛋白質組中的3694個代表蛋白的15199個肽抗原。由於抗體的多特異性，PETAL有可能與自然界中大量的蛋白質結合。利用PETAL mAbs製備抗體微陣列。利用不同蛋白質組的細胞裂解產物篩選PETAL陣列，驗證了mAbs用於蛋白質組規模蛋白靶向的可行性。已鑑定的抗體具有廣泛的應用前景，如用於人膜蛋白和核蛋白。從玉米和斑馬魚各自的蛋白質組特異性mAbs中，鑑定出了表型特異性mAb蛋白對。通過對正常與腫瘤細胞膜蛋白質組的差異篩選，證實缺失發現了候選治療靶點。以CD44v9為靶點的抗體被認為是構建肺鱗狀細胞癌（LUSCC）體內外抗體藥物結合物（ADC）的候選抗體。與當前情況相比，只需一小部分時間和成本就可產生高親和力的mAbs，PETAL能夠利用可用的基因組序列信息為蛋白質組生成親和力試劑並發現靶點。
結果
PETAL的免疫學基礎是抗體的多特異性（圖1A）。PETAL是一個由62208個mAbs組成的抗肽抗體庫，其目的是為不同蛋白質組的大量蛋白質提供結合物（圖1B）。當PETAL以陣列形式固定時（圖1C），它能夠產生蛋白質組規模的抗體並發現差異化靶標（圖1D）。
圖1抗體/靶標發現PETAL及其陣列的構建與應用
(A) Antibody multispecificity. An antibody bindsto an epitope/mimotope found within a variety of proteins from differentspecies, leading to high-affinity, specific binding of this antibody to a largenumber of proteins in nature. (B) PETAL construction. PETAL is a libraryof 62,208 mouse mAbs derived from immunization of more than 30,000 mice against15,199 diverse peptide antigens. PETAL has the potential to recognize a greatnumber of proteins in nature. (C) PETAL microarray construction. PETALis printed into an antibody microarray as a high-throughput platform forantibody/target discovery (left). Right panel shows the design/layout of thearray (red, visualized by a Cy5-conjugated anti-mouse antibody) and an arrayhybridization result using a protein sample (positive-binding mAb spots areshown as green). (D) Workflow for proteome-scale antibody generation andtarget discovery. Two array-screening applications are shown: direct screeningto identify proteome-specific mAbs and subsequent antibody applicationscreening and target identification or differential screening to find mAbs andtheir cellular targets associated with a specific phenotype.
PETAL的設計和構建
從代表418個蛋白質組的3694個蛋白質中共獲取了15199個肽抗原（稱為蛋白質組表位標籤或PETs）（圖1B、圖S1 A和表S1）。在每個蛋白質組中，使用最優啟發式blastp算法進行短肽序列比較，從蛋白質序列的獨特區域中選擇出PETs【23,24】。序列分析顯示PET序列具有隨機及多樣性（圖S1B）。
為了構建PETAL，合成了PET抗原，並仿照裝配線進行了大規模mAb開發操作來生成小鼠mAbs（圖1B）。使超過30,000隻小鼠免疫，平均一支PET免疫兩隻老鼠。共產生62208個小鼠mAbs（表S2）。每個雜交瘤細胞系用於製備腹水，腹水中含有1-20 mg的小鼠免疫球蛋白Gs（IgGs），其濃度從0.1-10 mg/ml不等；大多數在1~3 mg/ml範圍內。
PETAL多樣性
為了評估PETAL的多樣性，特別是使用同一PET產生的多個mAbs，對91個隨機選擇的雜交瘤進行了抗體V區測序，包括68個針對24個PETs的雜交瘤克隆，每個PET有2~5個mAbs，23個針對23個獨特PETs的雜交瘤（圖S1，C-E）。儘管發現兩個具有單一CDR胺基酸差異的抗體與不同的表位結合，認為互補性決定區（CDR）中胺基酸差異≥2的V區序列是「唯一的」【25】。接近90%的CDR序列是唯一的（圖S1D）。抗同一PET肽抗原的多個mAbs大多（80%~100%）是唯一的（圖S1E），這表明有效庫大小接近PETAL mAbs的總數，因為框架差異也可能導致不同的結合親和力，因此導致不同的特異性。
含62,208個mAbs的PETAL微陣列的構建
為了使用PETAL進行多路復用篩選，使用所有62208個抗體構建了抗體微陣列（圖1C）。使用與青色素5（Cy5）結合的抗小鼠二級抗體評估陣列列印質量（圖1C，左側兩幅小圖）。檢測到每點10~1000 pg（大部分是100~300 pg）的抗體，螢光強度在從500至60,000以上變化（圖S1F）。
通常，生物素化抗原，如10~100 μg蛋白質組樣本（圖1C，右側兩幅小圖）用於陣列篩選。使用三個獨立複製品中的人血漿樣本研究陣列篩選一致性。平均20000個抗體呈陽性結合[（信號強度-背景）/背景>3]，其變異係數（CV）為7%。超過90%的陣列陽性抗體具有500~10000的螢光強度。三次試驗的皮爾遜相關係數（r）為0.98（圖S1G）。使用這些數據建立了PETAL陣列，作為可重複篩選的平臺。
為了測試用於產生mAbs的庫/陣列，共使用81個重組蛋白篩選PETAL陣列（圖S2和表S3）。在酶聯免疫吸附試驗（ELISA）篩選陣列結合mAbs（抗原檢測限≤1 μg/ml）後，大約一半（47%；81中的38）的蛋白質平均每抗原產生3.7（141/38）個mAbs。免疫印跡試驗中檢測中，使用重組蛋白和內源性樣本，31%（81種中的25種）和26%蛋白質（81種中的21種）分別獲得成功（圖S2和表S3）。
PETAL靶向多種抗體蛋白質組及靶向發現
為了使用PETAL靶向廣泛及多樣的蛋白質組，本文使用了11個植物、動物和細菌的蛋白質組樣本來探測PETAL陣列（圖2）。每個蛋白質組的陣列陽性抗體數量為10000到20000。在螢光強度標度範圍內，每蛋白質組選擇約1000個陣列陽性抗體來探測內源性樣本，以估計適合免疫印跡的抗體數量（圖2A）。通常，20%到30%的抗體產生特異性（單一或主要的單一條帶）免疫印跡結果，提供2000到6000個潛在免疫印跡mAbs的蛋白質組。儘管蛋白質組結合mAbs在蛋白質組之間有明顯重疊（30~60%），但相同的抗體通常在不同的蛋白質組中特異性地識別不同大小的蛋白質（圖2B），可能是不相關的蛋白質。
圖2 PETAL靶向不同抗體蛋白質組和靶向發現
(A) Proteome-targeting PETAL mAbs forimmunoblotting. Successful rates (labeled as %) of immunoblotting (producingsingle/predominant single bands) were shown for 11 organisms by using a panelof ~1000 proteome-binding mAbs for each organism to probe proteome samples. Thetotal number of binding mAbs for a proteome was in the range of 10,000 to20,000. The specific tissues for immunoblotting were cow (breast, ovary, andliver), cotton (ovule), peach (leaf or fruit), grape (nuclear fraction ofseed), sugarcane (stalk), maize (seed), Pseudomonas aeruginosa (wholecell), silkworm (larva), zebrafish (embryo, heart or other tissues), axolotl(regenerating limb), and chicken (cell lines and tissues). (B) Differentproteins detected by the same mAb in two different proteomes. Four examples ofantibodies each recognized a specific band with different size in differentproteomes upon immunoblotting. (C) Identification of maize seeddevelopment stage–specific mAb-protein pairs. PETAL array screening using aproteome sample consisting of total protein extracts from maize seeds DAP3 andDAP17 produced a total of 12,427 binding mAbs. A selection of 1000 mAbs toprobe DAP3 and DAP17 samples yielded 206 class I mAbs (single or predominantsingle band) and additional 129 class II mAbs (multiple bands) uponimmunoblotting. Seventy differentially expressed mAbs between DAP3 and DAP17were used to IP their cellular binding proteins for MS analysis, resulting inthe identification of 19 proteins paired with 23 mAbs. Six proteins are shownin the right panel. Gel pictures from left to right show class I mAbimmunoblotting examples: silver staining (SS) of IP products and immunoblotting(IB) of input and IP products for Pb21831. (D) Identification of heartinjury–related proteins from zebrafish. From left to right: IF staining, silverstaining of IP products and immunoblotting of input and IP products forPb28030, and summary of the identified proteins.
利用蛋白質組特異性mAbs鑑定玉米籽粒發育和斑馬魚心臟再生的表型特異性MAbs和相應蛋白。利用玉米種子在授粉後第3天（DAP3）和DAP17（圖2C）兩個發育階段的蛋白質樣本，從1000個陣列陽性結合的mAbs中產生了一組335個免疫印跡陽性的mAbs【26】。其中包括206個單帶或優勢單一條帶mAbs（I類，21%；206/1000）和另外129個免疫印跡的多條（通常為2-3條）優勢條帶mAbs（II類）（圖2C）。在用Ⅰ類MAbs分別檢測DAP3和DAP17蛋白樣品後，利用免疫沉澱（IP）和液相色譜-串聯質譜（LC-MS/MS），選擇一組70個差異表達的靶點進行靶點鑑定。鑑定出19種與23個mAbs配對的蛋白質，圖2C中示出了6個例子，包括Sdh、Bt2和Sbe1，它們先前已被證明與玉米籽粒發育有關【26,27】。
在另一個例子中，通過使用損傷前和損傷後7天的心臟樣本進行IF分析，對通過篩選斑馬魚心臟蛋白質裂解物確定的一組45個結合mAbs進行進一步表徵（圖2D）。其中六個mAbs在受傷心臟中上調或下調。由這六種抗體結合的蛋白質通過IP和LC-MS/MS（圖2D，右）鑑定，包括具有已知心臟損傷保護功能的蛋白質 [例如，圖2D中的aldh2.2（左側和中間小圖）]【28】。
人膜蛋白和核蛋白的蛋白質組規模抗體生成
為了將PETAL應用於細胞器蛋白質組的抗體生成，利用人體PC9、HepG2、THP-1和Jurkat細胞繫膜蛋白和核蛋白的總蛋白提取物篩選PETAL陣列（圖3A）。每個樣本的陽性mAbs總數在12,000~18,000之間。
圖3人膜蛋白和核蛋白質組的蛋白質組規模抗體生成
(A) Protein identification for human membrane andnuclear proteome-specific PETAL mAbs. (B) An example (TFRC/CD71) foridentification of antibody binding protein. From left to right: SS for IPproduct, IB blot of input and after IP samples, coverage map of MS-identifiedpeptides, and IF and FACS data. The cell line for IF was selected according toHPA data. Membrane or nuclear proteins were labeled MEM or NUC. Negativecontrols (NC) for FACS included staining with blank and irrelevant IgG. (C)Examples of IF data for endogenously expressed membrane and nuclear proteins.ACTN4 and ACTP5B were stained under nonpermeable conditions. Other proteinswere stained under permeable staining conditions (also, see movies S1 to S6). (D)Proteins with two independent IP and immunoblotting mAbs. Panel label (SS andIB) was the same as in (B). Nuclear proteins were labeled in blue. (E)Abundance and function distribution of proteins identified from the Jurkat cellmembrane proteome. (F) Protein interactome example using Pb51585 againstPCCA. Protein-protein interacting map (right) analyzed by STRING with themass-identified proteins. (G) Snapshot of the Integrative GenomicsViewer showing sequencing read density of ChIP-seq data generated with antibodiesagainst SMRC1, SATB1, and NFIC in HepG2 cells. Chr1, chromosome 1.
To further screen for application-specific antibodies andto identify their cellular binding proteins, array-positive antibodies withhigh (>10,000), medium (2000 to 10,000), and low (500 to 2000) fluorescentintensity (fig. S3A) were selected. A total of 1439 positive antibodies formembrane proteomes and 379 for nuclear proteomes were subjected toimmunoblotting, IF/FACS, and IP assays (Fig. 3, B and C, fig. S3B, and movies S1 to S6), and their cellularbinding proteins were identified by LC-MS/MS (Fig. 3B). A total of 149 antibodies representing 107 proteinswere identified from membrane proteome screening (tables S4 and S5), includingknown CD and RAB (small guanosine triphosphatase) molecules CD3e, CD49d, CD71[transferrin receptor (TFRC)], CD222, CD5, CD2, CD44, RAB1B, and RAB14. Fornuclear proteomes, a total of 55 antibodies representing 42 proteins wereidentified (tables S4 and S5), including transcriptional regulators NONO, NFIC,TRIM28, CSNK2A1, MTA2, SATB1, SFPQ, and SMARCC1. About 20% of the proteins hadat least two independent antibodies that yielded similar IP and immunoblottingresults, strengthening the antibody validation quality (Fig. 3D) (29, 30). The successrate of target identification was consistent over a wide fluorescent intensityrange (fig. S3A), suggesting that more than 1000 proteins could be covered with12,000 to 18,000 proteome-binding antibodies.
只有20%到30%的陣列陽性抗體在免疫印跡分析中成功，這可能是由於篩選樣品中的天然蛋白質狀態所致。對於未成功免疫印跡數據的抗體（因此，其靶蛋白的大小未知），在IP和質譜之後，需要進行過表達或敲除實驗來確定其細胞結合蛋白。例如，通過將IF染色信號與過表達的PIEZO1-綠色螢光蛋白（GFP；圖S3C）共域化來確認抗體（Pb2795）對多通道離子通道PIEZO1的識別【31】。
為了研究抗體靶向蛋白質是否偏向於特定類型，對87個Jurkat蛋白質（67個膜蛋白和20個核蛋白；圖3E和圖S3D）的蛋白質豐度和功能類別[基因本體（GO）注釋]進行了檢測。根據PAXdb，這些蛋白質的豐度分布與Jurkat細胞中總膜或核蛋白的豐度分布相似【32】。根據HPA資料庫，由抗體鑑定的蛋白質來自不同的蛋白質家族和功能組，與人類總膜蛋白和核蛋白相似（圖3E和圖S3D），這表明抗體靶向的蛋白質種類廣泛，沒有明顯的偏倚。
為了測試抗體的附加應用，本文對選定的抗體進行了蛋白質複合物富集分析和染色質IP測序（ChIP序列）。蛋白質複合物富集分析結果如圖3F所示。例如，抗丙醯輔酶A羧化酶α（PCCA）的Pb51585可以根據基因/蛋白質相互作用檢索工具（STRING），拉低與先前映射的相互作用組類似的PCCA相互作用蛋白【33】。
在先前研究之後，在HepG2中進行了抗SMRC1、SATB1和NFIC抗體的ChIP測序分析【34】。共產生9390萬個測序結果，53.7%的測序結果被唯一地映射到人類參照基因組。這些結果被進一步處理並產生46380個峰，分別代表SMRC1、SATB1和NFIC的29441、3296和13643個結合位點（圖3G）。通過與商業ChIP抗體進行比較或利用ChIP定量聚合酶鏈反應（PCR）分析、結合位點分析和富集基序分析進行進一步驗證，均證實了這些轉錄因子的抗體可用於ChIP測序（圖S3、E至H）。
差分陣列篩選ADC抗體/靶標發現
ADC通過連接一種毒素，例如一甲基auristatin E（MMAE）和一種靶向內化腫瘤相關抗原（該抗原在腫瘤中的表達高於在正常組織中的表達）的抗體，選擇性地消除癌細胞【35】。為了確定適合ADC的候選靶點，本文用正常和腫瘤細胞蛋白質組篩選了PETAL陣列（圖4）。15000個肺膜蛋白組陽性抗體中的3000多個抗體在腫瘤和正常樣本之間的螢光強度倍數變化大於1.5（倍數變化大於1.1被認為具有顯著性，因為實驗重複的CV小於10%）（圖4A）。從500種抗體中篩選出4種抗體，其倍數變化範圍為1.5~5（圖4B和圖S4A），以顯示內化和間接細胞毒性。腫瘤組織內信號高2.4倍的抗體Pb44707被內化用於PC9細胞的細胞毒性研究，其中半內化時間為2.5 h，半數抑制濃度（IC50）為≤100 pm。IP和LC-MS/MS鑑定了CD44，其為一種假定的癌症幹細胞標記物，是最可能的靶蛋白（圖4C）【36】。通過肽和小幹擾RNA（siRNA）阻斷實驗進一步證實了靶蛋白，其中只有CD44v9特異性肽和CD44v靶向siRNA導致FACS表面螢光信號丟失（圖4、D和E）【37】。用FACS抗體滴定法測得Pb44707與PC9的細胞結合親和力[半數有效濃度（EC50）]為832 pM（圖S4B）【38】。
圖4 ADC治療靶點/抗體發現的差分陣列篩選
(A) Differential antibody screening for ADCtargets. Inset scatter plot shows differential distribution of antibody signalintensity of NSCLC and normal lung. n = 1. More than 3000tumor-high antibodies were identified. (B) An antibody candidate,Pb44707, for ADC. Antibody ID was labeled on the left of the IF image. IF (0and 4 hours) image (green, antibody; blue, DAPI) time course of normalizedsurface fluorescence in FACS and cell cytotoxicity data are shown. Internalizationhalf time (t1/2) and mean percent growth inhibition ± SEM (n =3) of the antibody is labeled. IF scale bar, 50 μm. (C) Pb44707 IP andMS. MS identified multiple peptides (sequence marked in red) from CD44s andpeptide sequences from variable region (shown as V number). (D) Peptideblocking of CD44v9. Graph of FACS analysis showing CD44v9 peptide but notCD44v6 peptide nor CD44s protein blocking binding of Pb44707 with PC9 cells. (E)siRNA targeting CD44v9 specifically decreases the FACS signal of Pb44707compared to control siRNA. Normalized MFI ± SEM of PC9 cells as detected byPb44707. (F) Representative images of the IHC staining of Pb44707 inLUSCC tumor tissue and paracancerous tissue from one patient. Scale bar, 300μm. (G) Images of the IHC staining of Pb44707 in representative vitalnormal human tissues. No expression of CD44v9 is detected in these tissues.Scale bar, 400 μm. (H) Quantification of total binding sites of Pb44707on the plasma membrane of a variety of tumor cell lines. (I) IC50 ofPb44707-ADC on tumor cell lines. n = 1 to 3 for different celllines, respectively. (J) Growth curves of the NCI-H1975 CDX tumors ofdifferent treatment groups (n = 7 per group). Treatment withAMT707, control ADC, or gefitinib (intraperitoneal injection, dosing once aday) was initiated 7 days after tumor inoculation and administered as indicatedby arrows. (K) Representative images of the IHC staining of Pb44707 in aLUSCC patient tumor tissue (top) and passage 1 (P1) PDX tumor tissue derivedfrom the same patient. Scale bar, 50 μm. (L) Growth curves of the PDXtumors of different treatment groups (n = 6 per group). Treatmentwith AMT707 or control ADC was initiated 35 days after tumor inoculation andadministered as indicated by arrows.
為探討CD44v9在腫瘤組織和正常組織中的表達，採用免疫組織化學（IHC）方法對Pb44707進行組織分析。CD44v9在60%的非小細胞肺癌（NSCLC）患者中明顯高表達，尤其是在接近90%的LUSCC（圖4F和圖S4，C至E）。此外，CD44v9的mRNA在LUSCC的各個階段都有較高的表達（圖S4F）。除皮膚外，大多數正常組織用Pb44707（圖4G）染色後呈陰性，在人群中呈高表達和低表達（圖S4、E和G）。為了將Pb44707構建入ADC分子，Pb44707與纈氨酸瓜氨酸（vc）-MMAE偶聯，平均藥物抗體比為4.23。Pb44707 MMAE（AMT707）對肺癌細胞系具有體外細胞毒性作用，其作用與CD44v9在肺癌細胞中的表達水平有關（圖4、H和I）。在細胞源性異種移植（CDX）模型和患者源性異種移植（PDX）模型中測試AMT707的體內療效。對於具有吉非替尼耐藥NSCLC細胞系NCI-H1975的異種移植模型（圖4J）而言，直至第47天和第57天，3和10 mg/kg劑量的AMT707分別完全抑制腫瘤生長，而對於治療對照組（即磷酸鹽緩衝鹽水（PBS）、吉非替尼和對照ADC），未產生效果或效果更低。在CD44v9陽性的LUSCC-PDX模型（圖4、K和L）中，AMT707以劑量依賴的方式抑制腫瘤生長，並且AMT707（10 mg/kg）完全抑制腫瘤生長。因此，通過陣列篩選，本文確定了多個內吞細胞表面靶點和至少一種抗體及其靶點，有可能構建一個ADC先導分子。
討論
為了實現蛋白質組規模的mAbs開發，本文構建了PETAL，一個由62208個mAbs及其相應的抗體陣列組成的雜交瘤庫，這是迄今為止報導的最大的抗體微陣列。利用抗體的多特異性，事實上PETAL可能含有大量蛋白質結合物【20-22】。結合PETAL和抗體微陣列的全局篩選能力，該平臺與現有方法相比，只花費一小部分時間和成本就可生成mAbs。適合的工作流程保證所選擇的mAbs具有較高親和力和理想特異性。本文已經證明了PETAL在大規模抗體產生、親和蛋白組學和治療靶點發現中的初步應用。
本文技術能夠靶向蛋白質組中的大量蛋白質（圖2），為蛋白質組/亞蛋白質組規模的抗體生成提供了解決方案【39】。本文證明，有效鑑定了靶向癌細胞和免疫細胞細胞表面蛋白和核蛋白的抗體。其中許多抗體非常適合IF和IP/ChIP。從1818個抗體的輸入中，鑑定出約200個能夠靶向149個獨立膜蛋白和核蛋白，進行免疫印跡/IF/FACS/IP的抗體。如果抗體/蛋白比率為12:1（1818/149），10000到20000個陣列陽性抗體將產生超過1000個結合蛋白。
基於抗體的功能蛋白質組學還不適用於非人類物種。本文證明了廣泛的蛋白質組樣本（包括目前為止經過測試的植物、動物和昆蟲）已經確定了與每個蛋白質組的大量蛋白質相對應的結合抗體。通過本文描述的蛋白質組樣品篩選和靶點鑑定，PETAL立即提供了免疫印跡mAbs，利用基因組測序信息對生物體內的蛋白質進行初步表徵。因此，進一步使用所鑑定的蛋白質組特異性mAb蛋白對從不同表型中探測樣本，以找到表型特異性mAb蛋白對。如文中所述，在斑馬魚心臟再生和玉米籽粒發育中發現了新的和以前報導的表型特異性蛋白質。對於所發現的每一種蛋白質，均獲得了能夠進行免疫印跡和IP的mAbs。抗體蛋白解卷積的總成功率與人膜/核篩選相似。因此，本文技術提供了一條簡單而有效的途徑，可以在目前沒有親和試劑的情況下對許多蛋白質組進行親和蛋白質組學研究。
當PETAL用於細胞膜蛋白質組的差異分析時，能為治療靶點的發現提供一種適用的方法。與其他功能基因組學（RNA幹擾或CRIPSR）或基於MS的方法相比，PETAL充分發揮了基於抗體的靶點發現的潛力，因為它同時傳遞了靶點和先導抗體。如本研究所示，通過比較腫瘤（NSCLC）與正常肺組織膜蛋白組的差異陣列篩選，確定了CD44v9靶向抗體，用於構建ADC分子AMT707。用IHC和FACS評價Pb44707的特異性。幾乎未觀察到非特異性結合。然而，在未來的臨床前研究中，特異性將通過一系列綜合手段得到進一步驗證。另一方面，由基於Pb44707的ADC引起的潛在皮膚毒性可以通過使用更溫和的有效載荷（如SN-38）來解決。當所有3000個「腫瘤高表達」抗體都經過篩選過程時，預計會發現更多的ADC候選抗體/靶點。
PETAL策略有一些局限性。PETAL的建造需要花費大量時間和精力，很難重複。此外，由於PETAL抗體是小鼠來源，因此進入治療前需要額外的抗體工程，包括人源化和免疫原性評估。然而，PETAL現在是一種預製資源，研究人員可以很容易地利用它來產生針對特定抗原或蛋白質組的mAbs，或者對蛋白質樣本進行差異篩選以確定表型特定的蛋白質靶點（圖S5）。儘管PETAL作為一個抗肽庫，但其它類型的抗原很可能產生同樣有用的抗體庫【5】。本文希望這項工作不僅可以作為一個即時資源，而且還可以激發新的雜交瘤庫，以供更多的應用。為此，本文建立的工業規模的mAb開發能力可以有效地構建數以萬計mAbs規模的雜交瘤庫/陣列。現有的PETAL只產生了用於20-30%的抗原的高親和力mAbs，通過增加庫的大小可以提高其應用，這反過來也可以提高PETAL mAbs的應用篩選和靶點鑑定的成功率。
總之，相比以前的抗體陣列和庫方法，PETAL有了顯著的改進。建立陣列篩選和抗體蛋白解卷積的工作流程，克服了以往方法成本高、開發時間長的缺點。本文期望PETAL通過支持蛋白質組規模的抗體生成和靶點分析來加速功能蛋白質組學。本文將努力為廣大科學界提供這一資源。本文相信PETAL可以激發其它抗體庫的製備，因此，該策略可以被許多其他研究者採用和探索。
MATERIALS AND METHODS
Peptide antigen selection
A total of 15,199 peptide antigens, called PETs, weredesigned from 3694 proteins representing 418 proteomes. Within each proteome,PETs were selected from unique regions of protein sequence using heuristicblastp algorithms optimal for short peptide sequence comparison (23, 29). Peptideantigens representative of predicted surface epitopes from a protein sequencewere selected (40, 41). Peptideswere mostly 10 to 12 amino acids in length to contain two to three potentialantibody epitopes (42). Predictedpeptide sequences with secondary structures including alpha helix and betasheet were omitted (43). Specialsequences including transmembrane motif, signal peptide, and posttranslationalmodification motif were also not selected. Only disordered or surface-loopedregions were selected. Hydrophobic peptides were not selected, and peptidehydrophilicity was calculated by the Hopp and Woods method (44). Last,peptides with more than one cysteine in the sequence were omitted to avoid synthesisdifficulties.
All the peptide antigens were chemically synthesized byGL Biochem (Shanghai) Ltd. The purity and molecular weight of each peptide wereevaluated with high-performance LC and MS.
Diversity analysis of PET library
The diversity of the PET peptide library (15,199peptides; see table S1 for detail) is evaluated by comparing the sequencesimilarity of all peptides against each other. The sequence identity (%)between a peptide and its closest homologs within the library was recorded. PETsequence similarity to two random peptide libraries generated computationallywas also compared. The first library was a collection of 15,199 peptidesequences randomly sampled from all species without considering amino acid preferencein different species (45). The secondlibrary was constructed by randomly sampling the entire human proteome (allconsensus coding sequences).
Construction of PETAL mAb library
MAbs were developed using a large-scale mAb developmentoperation modeled after an assembly line. In the antibody assembly line, eachof close to 100 highly trained technicians performs one to three discrete steps(for example, plating fusion cells onto 96-well plates or cell transfer from96- to 384-well plates) for making hybridomas. An internally built informaticsand data system (Antibody Assembler) is used for tracking materials and projectstatus. More than 90% of all materials used are bar-coded to minimize hand labeling.Many steps have automatic data analysis and decision making (for example, clonepicking). Together, antibody assembly line is scalable and cost efficient.PETAL is a premade library built by this highly efficient process. Aftertraditional hybridoma protocol (46) and theimmunization and fusion, a series of ELISA screens were performed using apeptide antigen titration from 1 × 10−7 to 1 × 10−10 Mto ensure that only the hybridoma clones with the highest affinity (forexample, able to detected antigen at a concentration less than 1 × 10−8 M)to peptide antigens were selected. IgG mAbs were selected using a Sigmaantibody isotyping kit (no. 11493027001). Four to six IgG hybridomas perpeptide antigen were selected for multiple rounds of limited dilutionsubcloning to ensure stability and monoclonality. Each hybridoma cell line wasused to prepare milliliters of ascites containing 1 to 10 mg of mouse IgGs.Mouse strains used for immunization and ascite production were BALB/c and F1from Shanghai Super-B&K Laboratory Animal Co. Ltd. The procedures for careand use of animals were approved by the Abmart Institutional Animal Care andUse Committee.
Hybridoma V-region sequencing
A mouse IgG primer set from Novagen (no. 69831-3) wasused to amplify the IgG variable region on antibody heavy chain (VH)and variable region on antibody light chain (VL) regions fromselected hybridoma clones. Briefly, 1 × 106 cells werecollected for each cell line. Total RNA was extracted using TRIzol reagent(Thermo Fisher, no. 15596026). The first-strand complementary DNA was amplifiedusing PrimeScript reverse transcription PCR kit from Takara (no. RR104A). PCRproducts with the expected size [an average size of about 400 base pairs (bp)for VH, and 360 bp for VL] weresequenced. The sequences of the PCR products were analyzed by IMGT/V-QUEST (www.imgt.org) (47) to definethe VH or VL regionsand the corresponding subelements. The uniqueness of antibody sequences wasevaluated by comparing full-length V (VH and VL),frame, or CDR sequences using clustal algorithm. The homology matrixes wereshown in the heat map format and that of the combined CDR sequences is shown infig. S1C as an example.
PETAL array construction and quality evaluation
Ascites of 62,208 PETAL mAbs were prepared in 162 384-wellplates and printed onto nitrocellulose-coated slides (FAST Maine Manufacturing,no. 10486111) in a high-density microarray format (named as PETAL array) usingthe Marathon System (Arrayjet Ltd., UK). Approximately 100 pl of ascites wasprinted for each antibody per spot. The array and block layout are shownin Fig. 1C. A total of 110 blocks were aligned into 10 rows and 11columns. Each block contains a subarray of 48 × 12 = 576 individual antibodyspots, except the subarrays in the last row were printed with 40 × 12 = 480(three blocks) or 39 × 12 = 368 (eight blocks) spots. Additional control rows,including a positioning fluorescent spot (Cy3) and a biotin–bovine serum albumin(BSA) gradient (0.4 to 50 pg) of eight spots, were also printed for each blocksimilar to previous antibody arrays (48, 49). Biotin-BSAwas prepared by saturated labeling of BSA with Thermo FisherSulfo-NHS-LC-Biotin (no. 21336) labeling reagent. PETAL arrays were stored at−80°C.
To evaluate PETAL array quality, the slides were blotteddirectly with a mixture of a Streptavidin-Cy3 (Sigma, no. S6402) and aCy5-labeled goat anti-mouse IgG (Jackson ImmunoResearch, no. 115-175-146), bothat a dilution of 1:3000 in 1× PBS. Fluorescence of Cy3 and Cy5 was recordedusing 532- and 635-nm channels by the GenePix 4200A Microarray Scanner(Molecular Devices LLC). Images were analyzed using GenePix Pro 6.0 software togive fluorescent intensities of each spot and its corresponding background.Missing or distorted spots, typically controlled under 5% of the total spots,were automatically marked by the software.
Reproducibility of array experiments was evaluated byincubating a triplicate of the same sample with three PETAL arrays. Thefluorescent intensity of each spot was normalized using biotin-labeled BSAsignal in each block and array. The normalized fluorescent intensities wereplotted between every experimental pair. Pearson product-moment correlationcoefficient value (R value) was calculated for each data pair inwhich the r value of +1 means total positive correlation and 0is no correlation (50).
Protein sample preparation
Cell lines [i.e., A431, A549, human embryonic kidney(HEK) 293 T, H1975, H226, Hela, HepG2, HL60, human umbilical cord endothelialcell (HUVEC), Jurkat, K562, MCF7, PC3, PC9, THP1, and U937] were purchased fromthe American Type Culture Collection (ATCC) or stem cell bank, Chinese Academyof Sciences (Shanghai, China). Cells were grown or maintained in Dulbecco’smodified Eagle’s medium (DMEM) or RPMI 1640 media following the ATCC cellculture guide. When cells were grown to ~80% confluence, they were dissociatedfrom culture plates by treatment with 1× PBS and 1 mM EDTA for 10 to 20 min.Trypsin was not used to avoid damages on cell surface proteins. Membrane ornuclear fractions of cell lysates were then prepared as described previously (51). Forwhole-cell lysis, 1× PBS containing 1% NP-40, 5 mM EDTA, and protease inhibitorcocktail (Calbiochem, no. 539134) was added to cells directly and incubated onice for 30 min. An ultrasonication step was performed before collecting thesupernatant. The prepared membrane fraction (MEM), nuclear fraction (NUC), andwhole-cell lysate (WCL) were labeled following the cell lines, respectively.The enrichments of marker proteins in MEM, NUC, and WCL fractions wereevaluated with anti-ATP5B (Abmart, no. M40013), histone 3.1 (Abmart, no.P30266), and β-tubulin (Abmart, no. M30109) mAbs, respectively.
For tissue samples from plants or animals, whole-celllysates were prepared following a protocol described previously (52). Briefly,frozen tissues were powdered in liquid nitrogen with a pestle, suspended in 10ml per 3 g of tissue extract protein extraction buffer [10 mM tris (pH 8.0),100 mM EDTA, 50 mM borax, 50 mM vitamin C, 1% Triton X-100, 2%2-mercaptoethanol, 30% sucrose] and incubated for 10 min. An equal volume oftris-HCl (pH 7.5)–saturated phenol was then added and vortex-mixed for 10 minat room temperature. The phenolic phase separated by centrifugation wasrecovered and reextracted twice with 10 ml of extraction buffer. Proteins inthe final phenolic phase were precipitated overnight at −20°C with 5× volumesof saturated ammonium acetate in methanol. Protein pellets collected bycentrifugation were washed twice with ice-cold methanol and once with ice-coldacetone. Pellets were then dried and dissolved with 500 mM triethylammoniumbicarbonate containing 0.5% SDS (pH 8.5). Bacteria lysates were prepared usingan ultrasonic apparatus.
Patient samples
All tissue microarray chips were purchased from ShanghaiOutdo Biotech Co. Ltd. Tumor and paracancerous tissues (normal) were freshlyexcised from a patient with NSCLC undergoing surgery. Tumor tissue and matchedparacancerous tissue were homogenized (53). Briefly,the specimens were cut into 0.5-mm sections before digestion with 0.1%collagenase IV (Gibco, no. 17104019) for 1 hour at 37°C. The cells were thenpassed through a 70-μm cell strainer (BD, no. 352350) and collected bycentrifugation for 15 min at 400g. Plasma membrane proteome extractswere prepared from single-cell suspensions of tissues.
Screening PETAL array with recombinant protein antigens
Recombinant protein antigens were first labeled withbiotin using EZ-Link NHS-LC-Biotin reagent (Thermo Fisher, no. 21366) and thenhybridized with the PETAL array. Array-bound proteins were incubated withStreptavidin-Cy3. The fluorescent intensity of mAb spots was then recorded bythe GenePix 4200A Microarray Scanner. Array-positive spots were defined as(signal-background)/background > 3. Protein-binding PETAL mAbs selected fromarray experiments were further screened through protein-mAb ELISA using adetection limit of 1 μg/ml of protein antigen. The ELISA-positive mAbs werethen validated on immunoblotting assays with recombinant proteins or endogenoussamples.
Screening PETAL array with proteomic antigens
Proteomic antigens including membrane, nuclear, orwhole-cell lysates were labeled with biotin using EZ-Link NHS-LC-Biotin reagentand incubated with the PETAL arrays. Following a similar procedure as describedabove, antibody spots positive in three independent experiments were thenranked by the averaged fluorescence intensities. A limited number ofarray-positive antibodies (1000 to 2000 according to the expected output ofeach screening) with high (>10,000), medium (2000 to 10,000), and low (500to 2000) fluorescent intensity were selected as candidate antibodies for furthervalidation assays.
Immunoblotting assays of PETAL mAbs
For recombinant protein immunoblotting, selected mAbswere used to probe 50, 10, 2, and 0.4 ng of recombinant protein antigens. Forimmunoblotting of endogenous human protein samples by PETAL mAbs, cell lineswere selected according to the protein expression profile from HPA and UniProtdatabases. For membrane or nuclear proteins, corresponding cellular fractionswere prepared for immunoblotting. Typically, 20 μg of protein was loaded ontoeach lane. Support-positive immunoblotting results were evaluated following thecriteria described by Antibodypedia (http://antibodypedia.com/text/validation_criteria#western_blot) and HPA (29). Basically,an antibody was qualified as immunoblotting positive when the size of a singleor predominant single band on immunoblotting matched or was within 10% of thepredicted antigen molecular weight. In some cases, an immunoblotting-positiveconclusion was enhanced when the same predicted protein band was detected intwo or more different cell lysates. Some antibodies detected multiple bandswith different sizes, but the predicted size protein band was also detected.
IF and FACS validation
A cell line with target protein expression (HPA data) wasselected for IF and FACS assays. Known/predicted subcellular localization ofthe target protein was also obtained from HPA or UniProt (table S4). For cellsurface proteins, IF and FACS assays were performed under nonpermeableconditions without detergent in the buffers. For intracellular proteins, thepermeable condition with 0.1% Triton added to the buffers was used throughout.Antibody binding signal was detected using Alexa Fluor 488 and 594 goatanti-mouse IgG secondary antibodies (Jackson ImmunoResearch, no. 115-545-003and no. 115-585-003). Briefly, cells attached on coverslips (IF assays) orsuspended in 1× PBS (FACS assays) were first fixed in 4% paraformaldehyde (PFA)for 10 min. PFA was then removed, and cells were rinsed three times with 1×PBS. Cells were blocked overnight at 4°C in blocking buffer (1× PBS containing10% normal goat serum, 0.1% Triton was added for intracellular proteins). Afterremoving the blocking buffer, cells were incubated with primary antibody(dilution in the blocking buffer at 1:100 to 1000) for 3 hours at roomtemperature. Cells were rinsed six times in 1× PBS before being incubated withfluorescence-labeled secondary antibody (diluted in blocking buffer at 1:500dilution ratio with 1:10,000 Hoechst 33258; Sigma, no. 94403) for 1 hour. Last,cells were rinsed three times with 1× PBS. IF images were recorded with a Nikonconfocal system A1Si. The three-dimensional reconstruction of the IF resultswas performed in ImageJ [National Center for Biotechnology Information, NIH(NCBI) free software]. IF staining patterns were compared with HPA data toconfirm the subcellular localization of the target proteins. The FACS data werecollected using a BD Accuri C6 Plus system. A control sample without primaryantibody and another sample with isotype control antibody were used.
IP and mass spectrum assays to identify antibody bindingprotein
IP assays were performed using cyanogen bromide(CNBr)–activated Sepharose 4B (GE, no. 17-0430-02) by following the user’smanual. Briefly, 200 μg of purified PETAL mAbs was cross-linked to 20 μl ofhydrolyzed CNBr beads and used to pull-down target protein from 1 mg of cellmembrane or nuclear protein samples. Typically, an excessive amount ofantibodies was used. A similar procedure was developed following theinstructions described previously (54) to identifythe binding proteins of the tested antibodies.
Essentially, target identification forimmunoblotting-successful (yielding single or predominant single band) mAbsused for IP was done by comparing the silver staining result of the IP productwith immunoblotting data on samples before and after IP. Expected size band(matched on silver staining and immunoblotting) was selected for MS analysis.For some mAbs, more than one band on SDS–polyacrylamide gel electrophoresis(PAGE) was selected for MS identification; several identified proteins could bethe binding targets of an antibody. For antibodies that failed in theimmunoblotting assay, their IP products separated on silver-stained SDS-PAGEwere compared to IP products from other mAbs. One predominant specific band orseveral stoichiometric-specific bands were selected for MS analysis.
Once one or more bands were selected for MS, 20 μl of IPproduct was separated in an SDS-PAGE gel and stained with Coomassie blue R-250.The selected bands were excised and sent to MS facilities (InstrumentalAnalysis Center of Shanghai JiaoTong University or Biological Mass SpectrometryFacility at Robert Wood Johnson Medical School and Rutgers, The StateUniversity of New Jersey) for target identification using LC-MS/MS on Thermo QExactive HF or Thermo Orbitrap-Velos Pro.
Mascot distiller (version 2.6, Matrix Science) or ProteinDiscovery software (version 2.2) was used to convert raw to mgf or mzML formatfor downstream analysis. The LC-MS/MS data were searched against UniProt human(557,992 proteins) for the human cell culture sample, UniProt zebrafish(61,675) for zebrafish tissue sample, or UniProt maize (137,106) for corntissue sample. Enzyme specificity was set as C terminal to Arg and Lys and allowedfor two missed cleavages. Furthermore, ±10 ppm and 0.02 Da (Thermo Q ExactiveHF) or 1 Da (Thermo Orbitrap-Velos Pro) were used as tolerance for precursor(MS) and product ions (MS/MS), respectively. Carbamidomethylated cysteine wasset as complete modifications. N-terminal protein acetylation and oxidation ofmethionine were set as potential modifications. Deamidation at asparagine andglutamine and oxidation at methionine and tryptophan were specified as variablemodifications.
To ensure MS data quality, we used a threshold of 20total identified peptide number or five nonredundant peptide number to achievehigh confidence of the identified protein. In analyzing the MS result for theantibody, the identified protein list was first prioritized using the totalidentified peptide number. Proteins that were identified in multiple differentantibodies were excluded. For most antibodies in this study, a unique proteinwith the highest total identified peptide number and matched protein sizedetected on the silver staining and immunoblotting was selected as the targetprotein.
The mass spectrometry proteomics data have been depositedto the ProteomeXchange Consortium via the PRoteomics IDEntifications (PRIDE) (55) partnerrepository with the dataset identifier PXD011629 (reviewer account details:username, reviewer41517@ebi.ac.uk; password, 7ZqfVOM8).
Abundance distribution and molecular function analysis
The identified membrane, nuclear, and other proteins werefrom the reference database (Nucleoplasm protein database and Nuclear membraneplus Plasma membrane protein database from HPA). Expression abundanceinformation of human proteins was obtained from the PAXdb. Functiondistributions were clustered using the PANTHER classification system (56) depending onthe molecular function.
ChIP-seq assay
The ChIP and input DNA libraries were prepared aspreviously described (34, 57). Briefly, 10million HepG2 cells were cross-linked with 1% formaldehyde for 10 min at roomtemperature and then quenched with 125 mM glycine. The chromatin was fragmentedand then immunoprecipitated with Protein A + G magnetic beads coupled withantibodies against SMRC1, SATB1, and NFIC. After reverse cross-linking, ChIPand input DNA fragments were used for library construction with NEBNext UltraLigation Module (NEB, no. E7445). The DNA libraries were amplified andsubjected to deep sequencing with an Illumina sequencer. The ChIP-seq dataprocessing was performed as we reported recently (57).Cis-regulatory sequence elements that mediate the binding of SMRC1, SATB1, orNFIC were predicted with MEME-ChIP (58).
Internalization assay
For the IF assay, live PC9 cells were cultured oncoverslips and incubated with mAbs (10 μg/ml) for 1 hour on ice before beingwashed three times with PBS. Cells were then cultured at 37°C for 0, 2, or 4hours before fixation with 4% PFA. Antibodies were then labeled with AlexaFluor 488–conjugated anti-mouse antibody. Images were acquired by Nikonconfocal system A1Si.
For the FACS assay, live PC9 cells were incubated withmAbs (10 μg/ml) for 0.5 hour on ice before being washed three times with PBS.Cells were then cultured at 37°C for up to 4 hours before fixation with 4% PFA.Cells were then stained with Alexa Fluor 488–conjugated anti-mouse antibody andanalyzed with FACS. Surface mean fluorescence intensity (MFI) was calculated.Surface mean fluorescent intensity (MFI), which represented surfacelocalization of mAbs, was measured by FACS.
Indirect cytotoxicity assay
PC9 cells were cultured in 96-well plates at 2000 perwell confluence overnight. Cells were treated with serial dilution of mAbstogether with MMAE-conjugated secondary goat anti-mouse IgG antibody (2 μg/ml)for 72 hours. Cell number was then calculated by Cell Counting Kit-8 (CCK8;Dojindo, no. CK04-20). Antibody-drug conjugation services were provided byLevena Biopharma, Nanjing.
In vivo tumor models
For the CDX model, 5 × 106 NCI-H1975cells were suspended in Matrigel (BD Biosciences, no. 354234) and injectedsubcutaneously to the right flank of female BALB/c nude mice (jsj-lab). Forstudies with the PDX model, the tumor fragments from patients with LUSCC werepassaged twice in nonobese diabetic–severe combined immunodeficient mice(Beijing Vital River Laboratory Animal Technology Co. Ltd.). Tumor fragmentsobtained from in vivo passage were then implanted subcutaneously in the rightflank of female BALB/c nude mice (jsj-lab). Body weight and tumor volume (0.5 ×length × width2) were measured every 3 days. Mice were randomizedinto control and treatment groups on the basis of the primary tumor sizes(median tumor volume of approximately 100 mm3). Pb44707-ADCs andcontrol ADCs were administered intravenously every third day and repeated for atotal of three times (Q3Dx3). Gefitinib (Selleck, ZD1839) was administeredintraperitoneally every day.
siRNA knockdown and overexpression
PC9 was transfected with siRNA targeting human CD44V9(sense, 5′-CUACUUUACUGGAAGGUUAtt-3′; antisense, 5′-UAACCUUCCAGUAAAGUAGtt-3′),which has been reported previously (37) or controlsiRNA (sense, 5′-UUCUCCGAACGUGUCACGUtt-3′; antisense,5′-ACGUGACACGUUCGGAGAAtt-3′) by Lipofectamine 2000 (Thermo Fisher, no.11668019) 48 hours before performing experiment.
PIEZO1-GFP plasmid used for overexpression validation wasa gift from D. Beech, which was described previously (59). PIEZO1-GFPwas transfected into HUVEC cells by Lipofectamine 2000. Cells were fixed andstained with anti-PIEZO1 antibody following IF procedure described above.
Peptide-blocking assay
Pb44707 (1 mg/ml) was preincubated with CD44 recombinantprotein (1 mg/ml; Abcam, no. ab173996) or CD44V peptide in 1:1 ratio at 4°Covernight before used in FACS analysis.
Antibody cellular binding site quantification
The antibody binding sites on cell lines were determinedwith the QIFIKIT (Dako, no. K0078) according to the manufacturer’s instructions.
SUPPLEMENTARY MATERIALS
Supplementary material for this article is availableat http://advances.sciencemag.org/cgi/content/full/6/11/eaax2271/DC1
Fig. S1. PETAL antigens, antibody diversity, and arrayperformance.
Fig. S2. PETAL screen by protein antigens.
Fig. S3. Proteome-scale antibody generation for humanmembrane and nuclear proteins.
Fig. S4. Differential array screening for ADC therapeutictarget/antibody.
Fig. S5. Two approaches to access PETAL library/array.
Table S1. Peptide antigens used for PETAL libraryconstruction.
Table S2. The list of the 62,208 antibodies and theirlocations on the microarray.
Table S3. Summary for PETAL screening on recombinantprotein targets.
Table S4. PETAL-screened protein targets from human celllines.
Table S5. Summary table for mAbs screened from human cellmembrane and nuclear extractions.
Movie S1. Movie showing cellular localization of TFRC inA431.
Movie S2. Movie showing cellular localization of ACTN4 inPC9.
Movie S3. Movie showing cellular localization of TIMM50in PC9.
Movie S4. Movie showing cellular localization of GOLIM4in PC9.
Movie S5. Movie showing cellular localization of NPM1 inA431.
Movie S6. Movie showing cellular localization of SFPQ inPC9.
Reference (60)
View/request a protocol for thispaper from Bio-protocol.
This is an open-access article distributed under theterms of the Creative CommonsAttribution-NonCommercial license, which permits use, distribution, and reproduction inany medium, so long as the resultant use is not for commercialadvantage and provided the original work is properly cited.
REFERENCES AND NOTES
1. K.-P. Koepfli, B. Paten; Genome10K Community of Scientists, S. J. O'Brien, The genome 10Kproject: A way forward. Annu. Rev. Anim. Biosci. 3, 57–111 (2015).
2. P. Carter, L. Smith, M. Ryan, Identificationand validation of cell surface antigens for antibody targeting inoncology. Endocr. Relat. Cancer 11, 659–687 (2004).
3. I. Hellstrom, H.J. Garrigues, U. Garrigues et al., Highly tumor-reactive,internalizing, mouse monoclonal antibodies to Le(y)-related cell surfaceantigens. Cancer Res. 50, 2183–2190 (1990).
4. R. S. Israeli, C. T. Powell, W.R. Fair et al., Molecular cloning of a complementary DNA encoding aprostate-specific membrane antigen. Cancer Res. 53, 227–230 (1993).
5. N. Hosen, Y. Matsunaga, K. Hasegawaet al., The activated conformation of integrin β7 is anovel multiple myeloma-specific target for CAR T cell therapy. Nat.Med. 23, 1436–1443(2017).
6. J. Benicky, R. Hafko, E. Sanchez-Lemuset al., Six commercially available angiotensin II AT1 receptor antibodiesare non-specific. Cell. Mol. Neurobiol. 32, 1353–1365 (2012).
7. M.C. Michel, T. Wieland, G. Tsujimoto, How reliableare G-protein-coupled receptor antibodies? Naunyn Schmiedebergs Arch.Pharmacol. 379, 385–388 (2009).
8. M. Uhlén, L. Fagerberg, B.M. Hallström et al., Tissue-based map of the humanproteome. Science 347, 1260419(2015).
9. P.J. Thul, L. Åkesson, M. Wiking et al., Asubcellular map of the human proteome. Science 356, eaa13321 (2017).
10. A. Jin, T. Ozawa, K. Tajiriet al., A rapid and efficient single-cell manipulation method forscreening antigen-specific antibody-secreting cells from human peripheralblood. Nat. Med. 15, 1088–1092 (2009).
11. M. DiCristina, L. Nunziangeli, M. A. Giubilei et al., Anantigen microarray immunoassay for multiplex screening of mouse monoclonalantibodies. Nat. Protoc. 5, 1932–1944 (2010).
12. S. T. Reddy, X. Ge, A.E. Miklos et al., Monoclonal antibodies isolated without screening byanalyzing the variable-gene repertoire of plasma cells. Nat.Biotechnol. 28, 965–969 (2010).
13. M. J. Feldhaus, R. W. Siegel, L.K. Opresko et al., Flow-cytometric isolation of human antibodies froma nonimmune Saccharomyces cerevisiae surface display library. Nat.Biotechnol. 21, 163–170 (2003).
14. B. R. Harvey, G. Georgiou, A. Hayhurstet al. , Anchored periplasmic expression, a versatile technology for theisolation of high-affinity antibodies from Escherichia coli-expressedlibraries. Proc. Natl. Acad. Sci. U.S.A. 101, 9193–9198 (2004).
15. C. Schaffitzel, J. Hanes, L. Jermutuset al., Ribosome display: An in vitro method for selection and evolutionof antibodies from libraries. J. Immunol. Methods 231, 119–135 (1999).
16. M. Mersmann, D. Meier, J. Mersmannet al., Towards proteome scale antibody selections using phage display.N.Biotechnol. 27, 118–128 (2010).
17. R. M. de Wildt, C. R. Mundy, B.D. Gorick et al., Antibody arrays for high-throughput screening ofantibody-antigen interactions. Nat. Biotechnol. 18, 989–994 (2000).↵
18. C. Wingren, Antibody-basedproteomics. Adv. Exp. Med. Biol. 926, 163–179 (2016).
19. H. Mao, J. J. Graziano, T. M.A. Chase et al.,Spatially addressed combinatorial protein libraries forrecombinant antibody discovery and optimization.Nat. Biotechnol. 28, 1195–1202 (2010).
20. R. A. Mariuzza, Multiple paths tomultispecificity. Immunity 24, 359–361 (2006).
21. L. C. James, P. Roversi, D.S. Tawfik, Antibody multispecificity mediated by conformational diversity.Science 299, 1362–1367 (2003).
22. A. Kramer, T. Keitel, K. Winkleret al., Molecular basis for the binding promiscuity of an anti-p24 (HIV-1)monoclonal antibody. Cell 91, 799–809 (1997).
23. S. F. Altschul, T. L. Madden, A.A. Schäffer et al., Gapped BLAST and PSI-BLAST: A new generation ofprotein database search programs. Nucleic Acids Res. 25, 3389–3402(1997).
24. L. Berglund, J. Andrade, J. Odeberget al., The epitope space of the human proteome. Protein Sci.17, 606–613 (2008).
25. E. A. Kabat, T. T. Wu, IdenticalV region amino acid sequences and segments of sequences in antibodies ofdifferent specificities. Relative contributions of VH and VL genes, minigenes,and complementarity-determining regions to binding of antibody-combiningsites. J. Immunol. 147, 1709–1719 (1991).
26. L. Jiang, X. Yu, X. Qi, Multigeneengineering of starch biosynthesis in maize endosperm increases the totalstarch content and the proportion of amylose. Transgenic Res. 22, 1133–1142 (2013).
27. S. M. de Sousa, M. del GiúdicePaniago, P. Arruda et al., Sugar levels modulate sorbitoldehydrogenase expression in maize. Plant Mol. Biol. 68, 203–213 (2008).
28. G.R. Budas, M.-H. Disatnik, D. Mochly-Rosen, Aldehydedehydrogenase 2 in cardiac protection: A new therapeutic target? TrendsCardiovasc. Med. 19, 158–164 (2009).
29. L. Berglund, E. Björling, P. Oksvoldet al., A genecentric Human Protein Atlas for expression profiles based onantibodies. Mol. Cell Proteomics 7, 2019–2027 (2008).
30. M. Uhlen, A. Bandrowski, S. Carret al., A proposal for validation of antibodies. Nat. Methods 13, 823–827(2016).
31. J. M.-J. Romac, R. A. Shahid, S.M. Swain et al., Piezo1 is a mechanically activated ion channel andmediates pressure induced pancreatitis. Nat. Commun. 9, 1715 (2018).
32. T. Geiger, A. Wehner, C. Schaabet al., Comparative proteomic analysis of eleven common cell lines revealsubiquitous but varying expression of most proteins. Mol. CellProteomics 11, M111.014050(2012).
33. D. Szklarczyk, J.H. Morris, H. Cook et al., The STRING database in 2017:Quality-controlled protein-protein association networks, made broadlyaccessible. Nucleic Acids Res. 45, D362–d368 (2017).
34. X. D. Zhao, X. Han, J.L. Chew et al., Whole-genome mapping of histone H3 Lys4 and 27trimethylations reveals distinct genomic compartments in human embryonic stemcells. Cell Stem Cell 1,286–298 (2007).
35. A. Beck, L. Goetsch, C. Dumontetet al., Strategies and challenges for the next generation of antibody-drugconjugates. Nat. Rev. Drug Discov. 16, 315–337 (2017).
36. V. Orian-Rousseau, CD44, a therapeutictarget for metastasising tumours. Eur. J. Cancer 46, 1271–1277(2010).
37. K. Kobayashi, H. Matsumoto, H. Matsuyamaet al., Clinical significance of CD44 variant 9 expression as a prognosticindicator in bladder cancer. Oncol. Rep. 36,2852–2860 (2016).
38. J. Willuda, L. Linden, H.G. Lerchen et al., Preclinical antitumor efficacy of BAY 1129980-anovel auristatin-based anti-C4.4A (LYPD3) antibody-drug conjugate for thetreatment of non-small cell lung cancer. Mol. Cancer Ther. 16, 893–904 (2017).
39. O. Stoevesandt, M.J. Taussig, Affinity proteomics: The role of specific bindingreagents in human proteome analysis. Expert Rev. Proteomics 9, 401–414 (2012).
40. T. P. Hopp, K.R. Woods, Prediction of protein antigenic determinants from aminoacid sequences. Proc. Natl. Acad. Sci. U.S.A. 78, 3824–3828 (1981).
41. G. W. Welling, W.J. Weijer, R. van der Zee et al., Prediction of sequentialantigenic regions in proteins. FEBS Lett. 188, 215–218 (1985).
42. E. D. Getzoff, J. A. Tainer, R.A. Lerner et al., The chemistry and mechanism of antibody binding toprotein antigens. Adv. Immunol. 43, 1–98 (1988).
43. P. Y. Chou, G.D. Fasma, Prediction of protein conformation. Biochemistry 13, 222–245 (1974).
44. T. P. Hopp, Use of hydrophilicity plottingprocedures to identify protein antigenic segments and other interactionsites. Methods Enzymol. 178, 571–585 (1989).
45. A. Moura, M.A. Savageau, R. Alves, Relative amino acid compositionsignatures of organisms and environments. PlOS One 8, e77319 (2013).
46. S. F. deStGroth, D. Scheidegger, Production of monoclonal antibodies:Strategy and tactics. J. Immunol. Methods 35, 1–21 (1980).
47. V. Giudicelli, X. Brochet, M.-P. Lefranc, IMGT/V-QUEST:IMGT standardized analysis of the immunoglobulin (IG) and T cell receptor (TR)nucleotide sequences. Cold Spring Harb. Protoc. 2011, 695–715 (2011).
48. P. Delfani, L. DexlinMellby, M. Nordström et al.,Technical advances of the recombinantantibody microarray technology platform for clinicalimmunoproteomics. PlOS One 11, e0159138 (2016).
49. C. A.K. Borrebaeck, C. Wingren, Design of high-density antibodymicroarrays for disease proteomics: key technological issues. J.Proteomics 72, 928–935 (2009).
50. J. Cohen, Statistical PowerAnalysis. Curr. Dir. Psychol. Sci. 1, 98–101 (2016).
51. J. V. Hagen, U. Michelsen, in Methods inenzymology, L. Jon, Ed. (Academic Press, 2013), vol. 533, pp. 25–30.
52. Y. Wang, W. Wang, J. Cai etal., Tomato nuclear proteome reveals the involvement of specific E2ubiquitin-conjugating enzymes in fruit ripening. Genome Biol. 15, 548 (2014).
53. D.-G. Zhang, A.-G. Jiang, H.-Y. Luet al., Isolation, cultivation and identification of human lungadenocarcinoma stem cells. Oncol. Lett. 9, 47–54 (2015).
54. E. Marcon, H. Jain, A. Bhattacharyaet al., Assessment of a method to characterize antibody selectivity andspecificity for use in immunoprecipitation. Nat. Methods 12, 725–731 (2015).
55. J.A. Vizcaíno, A. Csordas, N. del-Toro et al., 2016update of the PRIDE database and its related tools. Nucleic AcidsRes. 44, D447–D456 (2016).
56. H. Mi, A. Muruganujan, J.T. Casagrande et al., Large-scale gene function analysis with thePANTHER classification system. Nat. Protoc. 8, 1551–1566 (2013).
57. X. L. Zhang, J. Wu, J. Wanget al., Integrative epigenomic analysis reveals unique epigeneticsignatures involved in unipotency of mouse female germline stemcells. Genome Biol. 17, 162 (2016).
58. P. Machanick, T.L. Bailey, MEME-ChIP: Motif analysis of large DNAdatasets. Bioinformatics 27, 1696–1697 (2011).
59. J. Li, B. Hou, S. Tumova etal., Piezo1 integration of vascular architecture with physiologicalforce. Nature 515, 279–282 (2014).
60. J. Gertz, D. Savic, K.E. Varley et al., Distinct properties of cell-type-specific andshared transcription factor binding sites. Mol. Cell 52,25–36 (2013).