EED Maintains the Small Cell Lung Cancer Neuroendocrine Phenotype and Drives Lung Cancer Histological Transformation

EED Maintains the Small Cell Lung Cancer Neuroendocrine Phenotype and Drives Lung Cancer Histological Transformation | bioRxiv /* */ /* */ <!-- <!-- /*! * yepnope1.5.4 * (c) WTFPL, GPLv2 */ (function(a,b,c){function d(a){return"[object Function]"==o.call(a)}function e(a){return"string"==typeof a}function f(){}function g(a){return!a||"loaded"==a||"complete"==a||"uninitialized"==a}function h(){var a=p.shift();q=1,a?a.t?m(function(){("c"==a.t?B.injectCss:B.injectJs)(a.s,0,a.a,a.x,a.e,1)},0):(a(),h()):q=0}function i(a,c,d,e,f,i,j){function k(b){if(!o&&g(l.readyState)&&(u.r=o=1,!q&&h(),l.onload=l.onreadystatechange=null,b)){"img"!=a&&m(function(){t.removeChild(l)},50);for(var d in y[c])y[c].hasOwnProperty(d)&&y[c][d].onload()}}var j=j||B.errorTimeout,l=b.createElement(a),o=0,r=0,u={t:d,s:c,e:f,a:i,x:j};1===y[c]&&(r=1,y[c]=[]),"object"==a?l.data=c:(l.src=c,l.type=a),l.width=l.height="0",l.onerror=l.onload=l.onreadystatechange=function(){k.call(this,r)},p.splice(e,0,u),"img"!=a&&(r||2===y[c]?(t.insertBefore(l,s?null:n),m(k,j)):y[c].push(l))}function j(a,b,c,d,f){return q=0,b=b||"j",e(a)?i("c"==b?v:u,a,b,this.i++,c,d,f):(p.splice(this.i++,0,a),1==p.length&&h()),this}function k(){var a=B;return a.loader={load:j,i:0},a}var l=b.documentElement,m=a.setTimeout,n=b.getElementsByTagName("script")[0],o={}.toString,p=[],q=0,r="MozAppearance"in l.style,s=r&&!!b.createRange().compareNode,t=s?l:n.parentNode,l=a.opera&&"[object Opera]"==o.call(a.opera),l=!!b.attachEvent&&!l,u=r?"object":l?"script":"img",v=l?"script":u,w=Array.isArray||function(a){return"[object Array]"==o.call(a)},x=[],y={},z={timeout:function(a,b){return b.length&&(a.timeout=b[0]),a}},A,B;B=function(a){function b(a){var a=a.split("!"),b=x.length,c=a.pop(),d=a.length,c={url:c,origUrl:c,prefixes:a},e,f,g;for(f=0;f<d;f++)g=a[f].split("="),(e=z[g.shift()])&&(c=e(c,g));for(f=0;f<b;f++)c=x[f](c);return c}function g(a,e,f,g,h){var i=b(a),j=i.autoCallback;i.url.split(".").pop().split("?").shift(),i.bypass||(e&&(e=d(e)?e:e[a]||e[g]||e[a.split("/").pop().split("?")[0]]),i.instead?i.instead(a,e,f,g,h):(y[i.url]?i.noexec=!0:y[i.url]=1,f.load(i.url,i.forceCSS||!i.forceJS&&"css"==i.url.split(".").pop().split("?").shift()?"c":c,i.noexec,i.attrs,i.timeout),(d(e)||d(j))&&f.load(function(){k(),e&&e(i.origUrl,h,g),j&&j(i.origUrl,h,g),y[i.url]=2})))}function h(a,b){function c(a,c){if(a){if(e(a))c||(j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}),g(a,j,b,0,h);else if(Object(a)===a)for(n in m=function(){var b=0,c;for(c in a)a.hasOwnProperty(c)&&b++;return b}(),a)a.hasOwnProperty(n)&&(!c&&!--m&&(d(j)?j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}:j[n]=function(a){return function(){var b=[].slice.call(arguments);a&&a.apply(this,b),l()}}(k[n])),g(a[n],j,b,n,h))}else!c&&l()}var h=!!a.test,i=a.load||a.both,j=a.callback||f,k=j,l=a.complete||f,m,n;c(h?a.yep:a.nope,!!i),i&&c(i)}var i,j,l=this.yepnope.loader;if(e(a))g(a,0,l,0);else if(w(a))for(i=0;i (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0];var j=d.createElement(s);var dl=l!='dataLayer'?'&l='+l:'';j.src='//www.googletagmanager.com/gtm.js?id='+i+dl;j.type='text/javascript';j.async=true;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-M677548'); Skip to main content Home About Submit ALERTS / RSS Search for this keyword Advanced Search New Results EED Maintains the Small Cell Lung Cancer Neuroendocrine Phenotype and Drives Lung Cancer Histological Transformation Yixiang Li , Yasmin N. Laimon , Hyeonseo Cho , Marina Vivero , Gabriel Roberti De Oliveira , Andrew Delcea , Varunika Savla , Yuting Chen , Yavuz T. Durmaz , View ORCID Profile Xintao Qiu , Shweta Kukreja , Rong Li , Talal El Zarif , Wesley Lu , McKayla Van Orden , Jacob E. Berchuck , Roderick T. Bronson , Shuqiang Li , View ORCID Profile Hongbin Ji , Katerina Politi , Matthew L. Freedman , View ORCID Profile Henry W. Long , Sabina Signoretti , View ORCID Profile Matthew G. Oser doi: https://doi.org/10.1101/2025.07.07.663486 Yixiang Li 1 Department of Medical Oncology, Dana-Farber Cancer Institute and Brigham and Women’s Hospital, Harvard Medical School , Boston, MA 02215, USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site Yasmin N. Laimon 2 Department of Pathology, Brigham and Women’s Hospital, Harvard Medical School , MA 02115, USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site Hyeonseo Cho 1 Department of Medical Oncology, Dana-Farber Cancer Institute and Brigham and Women’s Hospital, Harvard Medical School , Boston, MA 02215, USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site Marina Vivero 2 Department of Pathology, Brigham and Women’s Hospital, Harvard Medical School , MA 02115, USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site Gabriel Roberti De Oliveira 2 Department of Pathology, Brigham and Women’s Hospital, Harvard Medical School , MA 02115, USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site Andrew Delcea 2 Department of Pathology, Brigham and Women’s Hospital, Harvard Medical School , MA 02115, USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site Varunika Savla 2 Department of Pathology, Brigham and Women’s Hospital, Harvard Medical School , MA 02115, USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site Yuting Chen 3 State Key Laboratory of Cell Biology, Shanghai Institute of Biochemistry and Cell Biology, Center for Excellence in Molecular Cell Science, Chinese Academy of Sciences , Shanghai 200031, China Find this author on Google Scholar Find this author on PubMed Search for this author on this site Yavuz T. Durmaz 1 Department of Medical Oncology, Dana-Farber Cancer Institute and Brigham and Women’s Hospital, Harvard Medical School , Boston, MA 02215, USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site Xintao Qiu 1 Department of Medical Oncology, Dana-Farber Cancer Institute and Brigham and Women’s Hospital, Harvard Medical School , Boston, MA 02215, USA 4 Center for Functional Cancer Epigenetics, Dana-Farber Cancer Institute , Boston, MA 02215, USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Xintao Qiu Shweta Kukreja 1 Department of Medical Oncology, Dana-Farber Cancer Institute and Brigham and Women’s Hospital, Harvard Medical School , Boston, MA 02215, USA 4 Center for Functional Cancer Epigenetics, Dana-Farber Cancer Institute , Boston, MA 02215, USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site Rong Li 1 Department of Medical Oncology, Dana-Farber Cancer Institute and Brigham and Women’s Hospital, Harvard Medical School , Boston, MA 02215, USA 4 Center for Functional Cancer Epigenetics, Dana-Farber Cancer Institute , Boston, MA 02215, USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site Talal El Zarif 1 Department of Medical Oncology, Dana-Farber Cancer Institute and Brigham and Women’s Hospital, Harvard Medical School , Boston, MA 02215, USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site Wesley Lu 1 Department of Medical Oncology, Dana-Farber Cancer Institute and Brigham and Women’s Hospital, Harvard Medical School , Boston, MA 02215, USA 5 Translational Immunogenomics Lab, Dana-Farber Cancer Institute , Boston, MA, USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site McKayla Van Orden 1 Department of Medical Oncology, Dana-Farber Cancer Institute and Brigham and Women’s Hospital, Harvard Medical School , Boston, MA 02215, USA 5 Translational Immunogenomics Lab, Dana-Farber Cancer Institute , Boston, MA, USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site Jacob E. Berchuck 6 Winship Cancer Institute, Emory University School of Medicine , Atlanta, GA 30322, USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site Roderick T. Bronson 7 Division of Immunology, Department of Microbiology and Immunobiology, Harvard Medical School , Boston, MA 02215 Find this author on Google Scholar Find this author on PubMed Search for this author on this site Shuqiang Li 1 Department of Medical Oncology, Dana-Farber Cancer Institute and Brigham and Women’s Hospital, Harvard Medical School , Boston, MA 02215, USA 5 Translational Immunogenomics Lab, Dana-Farber Cancer Institute , Boston, MA, USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site Hongbin Ji 3 State Key Laboratory of Cell Biology, Shanghai Institute of Biochemistry and Cell Biology, Center for Excellence in Molecular Cell Science, Chinese Academy of Sciences , Shanghai 200031, China 8 University of Chinese Academy of Sciences , Beijing 100049, China 9 School of Life Science, Hangzhou Institute for Advanced Study, University of Chinese Academy of Sciences , Hangzhou 310024, China 10 School of Life Science and Technology, Shanghai Tech University , Shanghai 200120, China Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Hongbin Ji Katerina Politi 11 Departments of Pathology and Internal Medicine (Section of Medical Oncology), Yale School of Medicine and Yale Cancer Center , New Haven, CT, USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site Matthew L. Freedman 1 Department of Medical Oncology, Dana-Farber Cancer Institute and Brigham and Women’s Hospital, Harvard Medical School , Boston, MA 02215, USA 4 Center for Functional Cancer Epigenetics, Dana-Farber Cancer Institute , Boston, MA 02215, USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site Henry W. Long 1 Department of Medical Oncology, Dana-Farber Cancer Institute and Brigham and Women’s Hospital, Harvard Medical School , Boston, MA 02215, USA 4 Center for Functional Cancer Epigenetics, Dana-Farber Cancer Institute , Boston, MA 02215, USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Henry W. Long Sabina Signoretti 2 Department of Pathology, Brigham and Women’s Hospital, Harvard Medical School , MA 02115, USA 12 Department of Oncologic Pathology, Dana-Farber Cancer Institute , Boston, MA Find this author on Google Scholar Find this author on PubMed Search for this author on this site Matthew G. Oser 1 Department of Medical Oncology, Dana-Farber Cancer Institute and Brigham and Women’s Hospital, Harvard Medical School , Boston, MA 02215, USA 13 Department of Medicine, Brigham and Women’s Hospital, Harvard Medical School , MA 02115, USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Matthew G. Oser For correspondence: Matthew_Oser{at}dfci.harvard.edu Abstract Full Text Info/History Metrics Supplementary material Preview PDF Abstract Lung cancer histological subtypes include lung adenocarcinoma (LUAD) and small cell lung cancer (SCLC). While typically distinct, combined LUAD/SCLC histology tumors occur, and LUAD can transform into SCLC as a resistance mechanism to targeted therapies, especially in EGFR -Mutant LUADs with RB1 / TP53 -inactivation. Although PRC2 complex expression increases during this transformation, its functional role has remained unclear. Using CRISPR-based autochthonous immunocompetent GEMMs, we demonstrate that inactivation of EED, the core PRC2 scaffolding subunit, impairs SCLC tumorigenesis and drives histological transformation from ASCL1-positive SCLC to LUAD through a transient NEUROD1-positive intermediate state. Mechanistically, EED loss de-represses bivalent genes co-marked by H3K27me3 and H3K4me3, including LUAD oncogenic RAS, PI3K, and MAPK pathway genes, to promote transformation to LUAD. Consistently, these same signaling genes are bivalently repressed in human SCLC patient-derived xenograft (PDX) tumors, suggesting a conserved PRC2-dependent mechanism to repress LUAD lineage oncogenic signaling to maintain the SCLC neuroendocrine identity. In a complementary EGFR -mutant LUAD GEMM with RB1/TP53 inactivation, EED was required for LUAD-to-SCLC transformation and distant metastasis upon EGFR withdrawal. These findings identify the PRC2 complex as a key epigenetic enforcer of SCLC neuroendocrine identity and nominate EED inhibition as a potential strategy to block SCLC transformation in high-risk LUAD. Introduction Lung cancer is the leading cause of cancer death worldwide and can be broadly divided into the distinct histological subtypes including non-small cell lung cancer (NSCLC) and small cell lung cancer (SCLC). NSCLC is further subdivided into lung adenocarcinoma (LUAD) and squamous cell carcinoma (SCC) which together account for ∼85% of all lung cancers, whereas SCLC accounts for the remaining ∼15% 1 . Although histological subtypes of lung cancer are often mutually exclusive, patients can present with combined LUAD, SCC, and SCLC histology tumors. This may arise from genetically distinct subclones or from lineage plasticity leading to histological transformation of one histology into another 2 , 3 . Lineage plasticity has been recognized in lung and other tumor types as a mechanism of therapeutic resistance, most notably in the context of: 1) Histological transformation of EGFR -Mutant LUAD to SCLC, as a mechanism of resistance to EGFR inhibitors 4 – 8 ; 2) Histological transformation of prostate adenocarcinoma (PRAD) to neuroendocrine prostate cancer (NEPC), as a mechanism of resistance to androgen deprivation therapy 3 , 9 . Both LUAD to SCLC and PRAD to NEPC histological transformation portend poor prognoses with relatively few therapeutic treatment options 3 . Although there have been several correlative studies identifying signatures associated with histological transformation from LUAD to SCLC 5 , causative functional drivers of histological transformation remain largely unknown. Recently, the first genetically-engineered mouse model (GEMM) of EGFR -Mutant LUAD to SCLC histological transformation was developed 10 . In this study, highly efficient histological transformation required: 1) Loss of Rb1 and Trp53 ; 2 tumor suppressors that are nearly universally lost in human SCLC; 2) Overexpression of non-degradable c-Myc; 3) Loss of the EGFR -Mutant oncogene 10 . Moreover, inactivation of the tumor suppressor Pten leading phosphoinositide 3-kinase (PI3K) pathway activation, when combined with Rb1 and Trp53 inactivation and c-Myc overexpression, allowed for SCLC tumorigenesis in alveolar type II cells 10 . This set of genomic alterations required for highly efficient EGFR -Mutant LUAD to SCLC histological transformation correlates with alterations that occur in the human counterpart validating the relevance of this GEMM 10 . For example, RB1 and TP53 loss near universally occurs in human SCLC 11 – 13 and commonly occurs in SCLC transformation 2 , 4 , 7 , 14 . Moreover, co-occurring RB1 and TP53 in EGFR -Mutant LUAD increase relative risk for SCLC transformation by 43-fold 14 . Alterations that lead to PI3K pathway activation including activation PIK3CA and PTEN are significantly associated with EGFR -Mutant LUAD to SCLC histological transformation in several independent studies 5 , 7 , 15 . Together these studies identify genomic alterations that drive LUAD to SCLC histological transformation and allow for the identification of high-risk patients; such as those with concurrent RB1 , TP53, and PIK3CA mutations 2 , 5 , 7 , 8 , 14 , 15 . However, causative functional drivers that constrain lung cancer histological subtypes and drive LUAD to SCLC histological transformation are not known. There are no current tractable therapeutic strategies to constrain lung cancer histological subtypes and block LUAD to SCLC histological transformation. SCLC exists broadly in four molecular subtypes: ASCL1, NEUROD1, POU2F3, and Inflammatory 16 , 17 . However apart from the POU2F3 molecular subtype which often homogenously expresses POU2F3 18 , there is increasing evidence for intra-tumoral heterogeneity of ASCL1 and NEUROD1 subtypes 19 and for ASCL1 to NEUROD1 subtype plasticity that can be driven by c-MYC 20 or the epigenetic modifier KDM6A 21 where c-MYC activation strongly drives further subtype plasticity toward a non-neuroendocrine (non-NE) subtype 20 . Moreover, a recent study identified SCLC subtypes marked by non-negative matrix factorization (NMF) clustering identifying NMF subsets associated with neuroendocrine states and immunotherapy response 12 . EED is the core scaffolding subunit of the PRC2 complex that canonically deposits tri-methylation marks on histone 3 lysine 27 (H3K27me3) to repress target genes 22 , 23 . Previous work demonstrated that SCLCs express high levels of EZH2 (the catalytic unit of PRC2) relative to LUADs 24 , and SCLC transformation is associated with signatures of high PRC2 activity 5 . Yet, the role of PRC2 in SCLC tumorigenesis—and whether this complex functionally enforces LUAD-to-SCLC histological transformation, a key mechanism of acquired resistance—remains unknown. Here, we leveraged CRISPR-engineered autochthonous immunocompetent GEMMs to dissect the role of the PRC2 complex in SCLC tumorigenesis and histological transformation, revealing that EED is essential for maintaining the neuroendocrine phenotype and functions as a gatekeeper of lung cancer lineage identity. Results EED Inactivation Drives Histological Transformation from SCLC to LUAD In Vivo To study the role of EED in SCLC tumorigenesis, we used a CRISPR/Cas9-based, autochthonous, immunocompetent GEMM of SCLC that we developed 21 , 25 – 28 where candidate target genes can be genetically inactivated at tumor initiation together with Rb1, Trp53, and Rbl2 ( RPR2 ) and the consequences of target gene inactivation during SCLC tumorigenesis can be studied. We generated adenoviruses that encode: 1) CMV-Cre to activate Cas9 expression; 2) sgRNAs targeting Rb1, Trp53, and Rbl2 ( RPR2 ); and 3) an sgRNA targeting Eed or a non-targeting sgRNA as a control ( Fig. 1A ). These adenoviruses could simultaneously inactivate all 4 targets and promote Cre expression in vitro ( Fig. S1A ). We then intratracheally (IT) injected these adenoviruses into the lungs of lox-stop-lox (LSL) Cas9-P2A-GFP mice for somatic CRISPR-based gene editing to make tumors that were genetically inactivated for Eed ( EED -Mutant RPR2 ) or were Eed WT ( EED -WT RPR2 ) ( Fig. 1A ). In this GEMM, all Cas9-positive tumor cells are GFP-positive. Eed was chosen rather than Ezh2 as EED is the scaffolding protein in all PRC2 complexes and hence eliminates the possibility of paralog compensation from EZH1 29 . Consistent with prior studies nominating EZH2 as a therapeutic target in SCLC 24 , 30 , Eed inactivation significantly delayed tumorigenesis, but tumors did eventually form in most mice injected with the EED -Mutant RPR2 adenovirus ( Fig. 1B-D ). Strikingly, the majority (70%) of (n=26 of 37 tumors) EED -Mutant RPR2 mice had histology consistent with pure LUAD and not SCLC ( Fig. 1C-D , S1B ). Another 27% of mice (n=10 of 37) had both SCLC and LUAD tumors with only 1 mouse showing pure SCLC ( Fig. 1C ). These findings reveal a dramatic phenotypic shift following EED loss. Download figure Open in new tab Figure 1. EED Inactivation Drives Histological Transformation from SCLC to LUAD In Vivo . A . Schematic of the adenovirus used for intratracheal injection (IT) into the lungs of lox-stop-lox (LSL)-Cas9 mice to generate autochthonous SCLC tumors that are Eed inactivated ( EED -Mutant) or Eed wild-type ( EED WT). RPR2 =sgRb1, sgTrp53, sgRbl2; sg “T”=sg Eed or sgControl (non-targeting sgRNA). B . Kaplan-Meier survival estimate of LSL-Cas9 mice IT injected with adenoviruses as indicated. Median overall survival: EED WT RPR2 : 44.93 weeks; EED -Mutant RPR2 : 65.00 weeks. N=6 EED WT RPR2 mice, n=9 EED -Mutant RPR2 mice. Log-rank (Mantel-Cox) test was used to calculate p value. C . Pie chart of lung cancer histology in EED -Mutant RPR2 tumors (n=37 tumors from 33 independent mice) and EED WT RPR2 tumors (n=11 tumors from 8 independent mice). OS=Osteosarcoma. D, E . H&E ( D ), and IHC staining for GFP, H3K27me3, EED, and Synaptophysin (SYP) ( E ) of lung tumors from representative EED WT RPR2 SCLC and EED -Mutant RPR2 LUAD mice. Scale bar is 1 mm ( D ) and 100 microns ( E ). Insets are 9X magnification. (F) Pie chart of liver metastases from EED -Mutant RPR2 (n=27 independent mice) and EED WT RPR2 (n=11 independent mice) mice. See also Figure S1. Immunohistochemistry (IHC) of tumors confirmed that all tumor cells were GFP-positive, had complete inactivation of EED and its methyltransferase product H3K27me3, and only SCLCs but not LUADs expressed the neuroendocrine marker Synaptophysin ( Fig. 1E , S1C ). CRISPR amplicon sequencing of select tumors similarly confirmed insertion-deletions (indels) in Eed and Rb1 leading to their genetic inactivation ( Fig. S1D-E ). The majority of EED -WT RPR2 SCLCs metastasized to the liver consistent with prior studies using the RPR2 SCLC model 26 , 31 and with SCLC being a highly metastatic tumor 1 ( Fig. 1F ) . In contrast, EED -Mutant RPR2 LUADs were confined to the lung without any evidence of distant metastasis ( Fig. 1F ) . Together, these data show that EED loss in vivo induces a striking histological transformation from SCLC to LUAD, suggesting strong selective pressure for lineage switching as a mechanism of escape from PRC2 loss. EED Inactivation Epigenetically Restores LUAD Oncogenic Signatures In Vivo To investigate the molecular consequences of EED loss during tumor lineage transformation, we performed RNA-sequencing on primary lung tumors from EED -Mutant RPR2 LUAD and EED -WT RPR2 SCLC mice. As expected, gene set enrichment using MSigDB 32 , ENCODE 33 , and ChEA 34 revealed that top upregulated genes in EED -Mutant tumors were targets normally bound and repressed by PRC2 complex members, including EZH2 and SUZ12 ( Fig. 2A–B ), consistent with functional PRC2 loss. Consistent with differences between human LUAD and human SCLC, EED -Mutant RPR2 LUAD tumors completely lost expression of canonical neuroendocrine (NE) markers including Chga , Insm1 , and Syp ( Fig. 2C ) and gained expression of genes involved in receptor tyrosine kinase (RTK) signaling, RAS signaling, and innate immune signaling including genes involved in antigen processing and presentation ( Fig. 2D ). Flow cytometry analysis of lung tumors validated that EED -Mutant RPR2 LUADs restored MHC class I antigen presentation and were significantly more infiltrated with immune and stromal cells ( Fig. S2A-D ). EED -Mutant RPR2 LUAD tumors correlated with both pure and pre-transformed human LUADs, while EED -WT RPR2 SCLCs correlated with both de novo and transformed SCLCs ( Fig. 2E ) supporting the translational relevance of this model 5 . Histologic analysis revealed that EED -Mutant LUADs displayed mucinous differentiation ( Fig. 2F ), a feature observed in subsets of human LUAD associated with KRAS mutations, NKX2-1 loss, and gastrointestinal lineage gene expression 35 – 38 . Correspondingly, EED -Mutant LUADs upregulated Cdx2 , Hnf4A , and Pdx1 ( Fig. 2G ), and were enriched for transcriptional signatures of gastrointestinal development ( Fig. 2H ) and mucinous LUADs ( Fig. 2I ). Download figure Open in new tab Figure 2. EED Inactivation Epigenetically Restores LUAD Oncogenic Signatures In Vivo . A . Principal component analysis (PCA) plot of RNA-seq data from 4 independent EED -WT RPR2 lung SCLC tumors (blue) and 4 independent EED -Mutant RPR2 LUAD lung tumors (red). B . Bar plot of ENCODE and ChEA Consensus Transcription Factors (TFs) enrichment score from RNA-seq data in A showing top enriched TFs whose targets were enriched EED -Mutant RPR2 LUAD vs. EED -WT RPR2 SCLC tumors. TFs with adjusted p value <0.05 were included. C . Heatmap depicting z scores of neuroendocrine marker expression from RNA-seq data in A . Dark blue is low, yellow is high. D . Dot plot of Gene Set Enrichment Analysis (GSEA) from RNA-seq data in A. FDR q-values are visualized by dot color, red=low, blue=high. NES=normalized enrichment score. Size of dot indicates percentage of genes enriched in the leading edge. E . Dot plot of the average correlation co-efficient between transcriptomic profiles of RPR2 tumors and human lung tumors with different histologies from Quintanol-Villalonga et al. 2021 5 . F . Representative H&E image of EED -Mutant RPR2 LUAD with mucinous features. Scale bar=100 microns. G . Heatmap of top 500 differentially upregulated and downregulated genes in EED -Mutant RPR2 LUAD vs EED -WT RPR2 SCLC ranked by z-score. Yellow is high, blue is low. H . Bar plot of negative log transformed adjusted p values of top organs from the mouse gene atlas whose transcriptomic profile enriched in differentially upregulated genes in EED -Mutant RPR2 LUAD vs. EED -WT RPR2 SCLC tumors with an adjusted p-value <0.05. I . Enrichment plot of human/mouse mucinous lung tumor signature from Guo et al. 2017 36 comparing EED -Mutant RPR2 LUAD vs EED -WT RPR2 SCLC. J, L . PCA plot of ATAC-seq ( J ) and H3K27ac ChIP-seq ( L ) from 3 independent EED -WT RPR2 SCLC lung tumors (blue) and 3 independent EED -Mutant RPR2 LUAD lung tumors (red). K, M . Heatmaps of differential binding peaks from ATAC-seq ( K ) and H3K27ac ChIP-seq ( M ) from lung tumor tissue in the indicated groups. N, O . Schematic ( N ) and volcano plot ( O ) showing differentially upregulated genes that are epigenetically activated (gain both accessibility and H3K27Ac) in EED -Mutant RPR2 LUAD vs. EED WT RPR2 SCLC. log2FC=log2 transformed fold change. P . Tracks of averaged RNA-seq, ATAC-seq, and H3K27ac ChIP-seq data across all samples above at Hnf4a (left), Met (middle), and Cdx2 (right) from the EED -WT RPR2 SCLC (blue) and EED -Mutant RPR2 LUAD (red) mouse tumors. Q, R . HOMER motif analysis showing top enriched motifs in EED -Mutant RPR2 LUAD from ATAC-seq ( Q ) and H3K27ac ChIP-seq ( R ). S. Schematic showing tumors escape EED inactivation by switching from SCLC to LUAD histology. Figure was made with BioRender. See also Figure S2. To define the chromatin-level changes underlying this reprogramming, we performed ATAC-seq and H3K27ac ChIP-seq on EED -Mutant LUADs and EED -WT SCLCs ( Fig. 2J-M , S2E-F ). Integrated analysis revealed that ∼80% of upregulated genes gained both chromatin accessibility and H3K27 acetylation ( Fig. 2N ), indicating widespread epigenetic activation. These included transcription factors associated with gastrointestinal differentiation such as Hnf4a and Cdx2 and LUAD oncogenes including Met ( Fig. 2O-P ). Motifs enriched in EED -Mutant RPR2 LUAD for both chromatin accessibility and H3K27ac also included HNF4A ( Fig. 2Q-R ). Together, these findings indicate that EED loss ultimately drives histological transformation from SCLC to LUAD in vivo associated with epigenetic activation of LUAD oncogenic programs and gastrointestinal differentiation factors ( Fig. 2S ) . EED Inactivation Promotes Rare NEUROD1-positive SCLC with Early Features of LUAD Oncogenic Signaling While EED -Mutant RPR2 mice predominantly developed LUAD tumors, a small subset formed tumors with SCLC histology ( Fig. 1C ). These rare EED -Mutant RPR2 SCLCs exhibited distinctive morphology, including giant cells and enlarged nuclei compared to EED -WT RPR2 SCLCs ( Fig. 3A ), and were transcriptionally distinct, with enrichment for EMT, inflammatory response, and pancreatic beta cell signatures ( Fig. 3B–C ). The latter included Neurod1 , a transcription factor that marks the NEUROD1-high (SCLC-N) molecular subtype of human SCLC 12 , 16 , 17 . These tumors showed high expression of Neurod1 and Neurod1 -correlated genes, alongside reduced expression of Ascl1 and its transcriptional program ( Fig. 3D–E ), suggesting that EED inactivation promotes ASCL1 to NEUROD1 subtype switching. Consistently, EED -Mutant RPR2 SCLCs correlated with human SCLC-N gene expression profiles from IMpower133 12 ( Fig. 3E ). Download figure Open in new tab Figure 3. EED Inactivation Promotes Rare NEUROD1-Positive SCLC with Early Features of LUAD Oncogenic Signaling. A . Representative H&Es of an EED -WT RPR2 SCLC and an EED -Mutant RPR2 SCLC. Yellow arrows point to giant-cell feature in the EED -Mutant RPR2 SCLC. Scale bar=100 microns. B . PCA plot of RNA-seq data from EED -WT and EED -Mutant RPR2 SCLC and LUAD tumors. C . Dot plot of hallmark pathway GSEA output comparing EED -Mutant RPR2 SCLCs vs. EED -WT RPR2 SCLCs from B. Dot color indicates FDR q-values. Red=low, blue=high. Size of dot indicates percentage of genes enriched in the leading edge. NES=normalized enrichment score. Leading edge was visualized by the size of the dots. EMT=epithelial to mesenchymal transition. D . Heatmap showing neuroendocrine marker expression from z-scored RNA-seq data in B. Dark blue is low, yellow is high. E . Dot plot of the indicated gene sets from RNA-seq data in B comparing EED -Mutant RPR2 SCLCs vs. EED -WT RPR2 SCLCs. FDR q-values are visualized by dot color, red=low, blue=high. NES=normalized enrichment score. Size of dot indicates percentage of genes enriched in the leading edge. F,G. Heatmaps of differential binding peaks from ATAC-seq ( F ) and H3K27ac ChIP-seq ( G ) comparing 2 independent EED -Mutant RPR2 SCLC vs. 3 independent EED WT RPR2 SCLC tumors. H, I . Schematic ( H ) and volcano plot ( I ) showing differentially upregulated genes that are epigenetically activated (gain both accessibility and H3K27Ac) in EED -Mutant RPR2 SCLC vs. EED WT RPR2 SCLC. log2FC=log2 transformed fold change. J . Tracks of averaged RNA-seq, ATAC-seq, and H3K27ac ChIP-seq data across all samples above from the EED -WT SCLC (blue), EED -Mutant SCLC (yellow) and EED -Mutant LUAD (red) mouse tumors at Neurod1, Isl1, Met , and Fgfr1 . K, L . Immunoblot analysis ( K ) and volcano plot of RNA-seq data ( L ) of 1014 RPR2 EED-WT SCLC cells transduced with sgRNAs targeting EED or a non-targeting sgRNA as a control (sgControl). For N, vertical lines indicated fold change of 2 (absolute value). Horizontal line, adjusted p value=0.05. M . Immunoblot analysis of 1014 RPR2 EED-WT SCLC cells treated with the EZH1/2 or EED inhibitors indicated. MAK683 (EED inhibitor)=3 uM, Tulmimetostat (EZH1/2 inhibitor)=100 nM, Valemetostat (EZH1/2 inhibitor)=100 nM, Tazemetostat (EZH2 inhibitor)=5 micromolar. N . Schematic of overlap between EED ChIP-seq and H3K27me3 ChIP-seq from EED WT RPR2 SCLC tumors to identify direct EED target genes. O . Schematic showing strategy to integrate direct EED target genes with genes reactivated upon EED loss using RNA-seq, ATAC-seq and H3K27ac ChIP-seq to identify direct functional EED targets in EED -Mutant SCLC. P . Rank plots of RNA-seq log2FoldChange of functional EED targets in EED -Mutant SCLC. Q. Dot plots of hallmark pathway enrichment analysis of direct functional EED targets in EED -Mutant SCLC. EMT: Epithelial to mesenchymal transition. BH adjusted p values were visualized by the dot color, red=low, blue=high. Enriched gene counts were visualized by the size of the dots. R . Gene set enrichment score of direct functional EED targets in EED -Mutant SCLC ( O ) on different subtypes of human SCLC patient samples RNA-seq from IMPOWER133 (Nabet et al., 2024 12 ). SCLC-N=NEUROD1+ SCLC, SCLC-A=ASCL1+ SCLC, SCLC-I-NE=Inflammatory Neuroendocrine SCLC, SCLC-I-non-NE=Inflammatory non-Neuroendocrine SCLC. One way ANOVA test with multiple comparisons was used. *=p<0.05, **=p<0.01, ***=p<0.001, ****=p<0.0001. See also Figure S3. Similarly, ATAC-seq and H3K27ac ChIP-seq also revealed that top enriched motifs with increased chromatin accessibility in EED -Mutant RPR2 SCLCs were NEUROD1/NEUROG3 and ISL1/2 motifs ( Fig. 3F-G , S3A-G ) with Neurod1 and Isl1 gaining both chromatin accessibility and H3K27ac ( Fig. 3H-J ). Interestingly, genes associated with LUAD oncogenic RTK signaling including Met and Fgfr1 were also significantly epigenetically activated in EED -Mutant RPR2 SCLCs relative to EED -WT RPR2 SCLCs ( Fig. 3H-J ). EED genetic inactivation in an ASCL1-positive murine SCLC cell line (1014) similarly upregulated Neurod1 , Met , and Fgfr1 mRNA ( Fig. 3K–L ). Moreover, EED inactivation or treatment with EZH1/2 or EED inhibitors increased NEUROD1 protein expression ( Fig. 3K , 3M ). RNA-seq also revealed an EED -Mutant RPR2 tumor that was positive for POU2F3 which marks a non-NE SCLC subtype 17 ( Fig. S4A-C ). However, EED inactivation in 1014 cells did not induce POU2F3 expression ( Fig. S4D ) suggesting plasticity toward POU2F3-positive SCLCs after EED inactivation is likely indirect. Finally, ChIP-seq for EED and H3K27me3 identified 6,072 direct EED target genes, 289 of which were epigenetically activated and expressed in EED -Mutant RPR2 SCLCs ( Fig. 3N–O , Fig. S3H ), including Neurod1 , Isl1 , and Pdx1 ( Fig. 3P ). Gene signatures enriched in these tumors were associated with cell state changes and NEUROD1 subtype switching including epithelial mesenchymal transition (EMT), pancreas beta cells, and the inflammatory response ( Fig. 3Q ). These genes were also highly enriched in the NEUROD1 and inflammatory human SCLC subtypes ( Fig. 3R ) validating relevance of EED -Mutant RPR2 SCLC tumors to the human counterpart. Together, these results show that EED inactivation de-represses NEUROD1 and promotes rare NEUROD1-positive SCLC tumors with early transcriptional activation of LUAD oncogenic pathways, suggesting a transition state on the path to LUAD histological transformation. EED Inactivation Promotes SCLC-to-LUAD Transformation Through a NEUROD1-Positive Intermediate Cell State and Requires Cues from the Tumor Immune Microenvironment In Vivo Although EED loss promoted NEUROD1 expression in vitro and generated occasional NEUROD1+ SCLC tumors in vivo , the vast majority of tumors in EED -Mutant RPR2 mice were LUADs ( Fig. 1C–D ), suggesting selective pressure favoring LUAD fate. Consistently, careful histological examination revealed rare early NE-positive lesions within LUAD-bearing lungs that expressed various SCLC NE subtype transcription factors including ASCL1, NEUROD1, and POU2F3 ( Fig. S4E ). To further explore the relationship of NEUROD1 to lung cancer histological subtypes formed after EED inactivation in vivo , we first performed IHC for NEUROD1 and ASCL1 ( Fig. 4A ). Consistent with our results above, EED -Mutant RPR2 SCLC tumors expressed NEUROD1, while EED -WT RPR2 tumors expressed ASCL1 ( Fig. 4A ). Surprisingly, EED -Mutant RPR2 LUADs retained sporadic NEUROD1-positive cells intermixed within tumors ( Fig. 4A-B ). This was a unique feature of EED -Mutant RPR2 LUADs as no NEUROD1-positive cells were found in a de novo LUAD EGFR -Mutant RPR2 GEMM (see Methods) nor in EED -WT RPR2 SCLCs ( Fig. 4A-B ). Neurod1 mRNA expression was also significantly elevated in EED -Mutant RPR2 LUADs relative to EED -WT RPR2 SCLCs ( Fig. 4C ). These data suggest that EED -Mutant RPR2 LUADs may have evolved from NEUROD1-positive EED -Mutant RPR2 SCLCs. Download figure Open in new tab Figure 4. EED Inactivation Promotes SCLC-to-LUAD Transformation Through a NEUROD1-Positive Intermediate Cell State and Requires Cues from the Tumor Immune Microenvironment In Vivo . A . Representative H&E and IHC staining for ASCL1, NEUROD1, H3K27me3 and EED of lung tumors from GEMMs of EED -WT RPR2 SCLC, EED -Mutant RPR2 SCLC, EED -Mutant RPR2 LUAD, and de novo LUAD. Red arrows indicate rare scattered NEUROD1 positive staining. Scale bar=100 microns. Insets are 9X magnification. B . Quantitation of NEUROD1-positive cells from NEUROD1 IHC in A. N=4 EED WT SCLC RPR2 tumors, n=2 EED -Mutant SCLC RPR2 tumors, n=10 EED -Mutant LUAD RPR2 tumors, n=2 de novo LUAD tumors. C . Dot plot of normalized Neurod1 mRNA expression from RT-qPCR in GEMM tumors indicated. N= 2 EED WT SCLC RPR2 tumors, n=2 EED -Mutant SCLC RPR2 tumors, n=39 EED -Mutant LUAD RPR2 tumors, n=2 de novo LUAD tumors. D, E . Uniform Manifold Approximation and Projection (UMAP) of all cells ( D ) and tumor cells ( E ) from single-nucleus RNA-seq (snRNA-seq) from 3 EED -Mutant RPR2 LUAD tumors from 3 independent mice. N=27,117 total cells, n=16,590 tumor cells. F . Dot plot showing average expression of indicated genes in each tumor cell cluster. Dot size represents percentage of cells expressing the indicated genes. Dot color represents expression level, blue=high, white=low. G . Phylogenetic tree of tumor cell clusters from EED -Mutant RPR2 LUAD tumor 1350 from E using Maximum Parsimony method. H . Heatmap of normalized expression level of top 300 trajectory-defining genes along the EED -Mutant RPR2 LUAD tumor cell trajectory. Blue is low, red is high. I. UMAP of all cells from single-cell RNA-seq (scRNA-seq) on 3 independent EED -WT SCLC (n=7889 cells), 2 EED -Mutant SCLC (n=2064 cells), and 4 EED -Mutant LUAD lung tumors (n=2013 cells). Each tumor is from an independent mouse. Mϕ: macrophages. DC: dendritic cell. J-L . Schematic ( J ), immunoblot analysis ( K ), and RT-qPCR ( L ) of 1014 RPR2 EED -WT SCLC cells transduced with an sgRNA targeting EED (sg Eed ) or a non-targeting sgRNA as a control (sgControl) and then subcutaneously implanted in immunodeficient (NCr nude) or immunocompetent (BL6J) mice. M . Heatmap of both tumor and immune cell types indicated comparing EED -Mutant LUAD, EED -Mutant SCLC, and EED -WT SCLC from the scRNA-seq data in I . N,O . UMAP ( N ) and quantification ( O ) of percentages of two M2-like Mϕ clusters relative to total cells in each sample. P,Q . Schematic ( P ) and rank plot ( Q ) of scaled receptor expression fold changes in EED -Mutant LUAD tumor cells whose cell-cell interactions were enriched in EED -Mutant LUAD tumors when compared to EED WT SCLC tumors and receptors upregulated after EED genetic inactivation in 1014 EED -WT SCLC cell line after EED CRISPR-mediated inactivation in vitro . For Met , the corresponding ligand ( Hgf ) and the tumor microenvironment component with significantly higher expression (Mϕ) is indicated in parenthesis. Mϕ: macrophages. R . Schematic showing SCLC to LUAD histological transformation occurs through a NEUROD1-positive transient intermediate state that requires cues from the tumor immune microenvironment. Figure was made with BioRender. *=p<0.05, **=p<0.01, ***=p<0.001, ****=p<0.0001. Student’s t test was used to calculate two-sided p-value. See also Figure S4, S5. To investigate this possibility, we performed single-nucleus RNA-seq (snRNA-seq) on three independent EED -Mutant RPR2 LUADs. Unsupervised clustering identified four tumor cell clusters ( Fig. 4D–E ). One cluster (cluster 3) expressed NE markers including Neurod1 , Chga , and Ncam1 and accounted for ∼4% of tumor cells, while other clusters expressed LUAD genes such as Met ( Fig. 4F ). To ask whether there was a clonal relationship between cluster 3 Neurod1 -positive cells with the other tumor cell clusters, we used InferCNV 39 to estimate copy number variations (CNVs) of each cluster within samples. InferCNV revealed diverse CNV landscapes between samples ( Fig. S5A-C ). However, we observed substantial shared copy number alterations between Neurod1 -positive cluster 3 tumor cells and tumor cells within other clusters ( Fig. S5A-C ) suggesting tumor cells from the same sample are likely derived from a common clone. Neurod1 -positive cluster 3 tumor cells showed significantly less CNV events compared to other clusters within the same tumor where tumor cells of other clusters gained or lost CNVs relative to Neurod1 -positive cluster 3 cells ( Fig. S5D ). Indeed, phylogeny tree construction also supported Neurod1 -positive cluster 3 tumor cells as the likely ancestor ( Fig. 4G , S5E-F ). Similarly, pseudo-time cell trajectory analysis revealed that first NE genes ( Ascl1 , Neurod1 , Ncam1 , Chgb ) were highly expressed followed by persistence of Neurod1 and Ncam1 expression with eventual loss of NE genes and upregulation of LUAD and GI differentiation genes ( Met , Hnf4a , Nr1h4 ) ( Fig. 4H , S5G-H ). Interestingly, EMT genes including Cdh6, Clu , and Itga2, were also highly expressed during the persistent Neurod1+ and Ncam1+ intermediate cell state. To further validate this trajectory, we performed single-cell RNA-seq (scRNA-seq) on EED -Mutant RPR2 LUADs (n=4) and EED -Mutant RPR2 SCLCs (n=2), comparing them to previously published scRNA-seq data from EED -WT RPR2 SCLCs (n=3) 21 ( Fig. 4I , S5I ). As expected, EED -Mutant SCLCs expressed Neurod1 , Isl1 , and other NE markers, while EED -Mutant LUADs lost NE features and expressed Hnf4a and Muc13 ( Fig. S5J ). Unsupervised ordering of EED -WT and EED -Mutant tumor cells predicted a single lineage trajectory of EED -WT RPR2 SCLC (ASCL1-positive) to EED -Mutant RPR2 SCLC (NEUROD1-positive) to EED -Mutant RPR2 LUAD (HNF4A-positive), further supporting NEUROD1-positive SCLC is an intermediate state during histological transformation from SCLC to LUAD ( Fig. S5K-M ). Despite upregulating Neurod1 and Met , EED inactivation in 1014 cells in vitro did not induce LUAD lineage markers (e.g. Hnf4a ) or complete transformation, suggesting that extrinsic in vivo cues from the tumor immune microenvironment are required to drive SCLC-to-LUAD transition ( Fig. 3L ). To test this, we transplanted EED -isogenic 1014 cells into immunocompetent (BL6J) and immunodeficient (NCr nude) mice ( Fig. 4J ). In immunodeficient hosts, EED -inactivated tumors retained high NEUROD1 expression and failed to activate MAPK signaling. In contrast, EED -inactivated tumors in immunocompetent mice lost NEUROD1 and robustly activated MAPK signaling and Met , suggesting that the tumor immune microenvironment promotes LUAD fate selection following EED loss ( Fig. 4K-L ). To explore which immune-derived cues might promote LUAD histological transformation, we analyzed the tumor immune microenvironment (TIME) ( Fig. 4I ) . Consistent with differences between human SCLC and human LUAD, EED -WT RPR2 SCLCs were mostly comprised of tumor cells with relatively few immune cells, while EED -Mutant RPR2 LUADs were significantly more immune cell infiltrated ( Fig. 4M ). Interestingly, EED -Mutant RPR2 SCLCs also had significantly more immune cell infiltration relative to EED -WT RPR2 SCLCs ( Fig. 4M ). Immunosuppressive M2-like macrophages 40 , 41 expressing Msr1 , Arg1 , and Tgfb1 were enriched in both EED -Mutant LUADs and SCLCs ( Fig. 4N-O , S5N-O ). To identify potential microenvironmental drivers of transformation, we performed a cell–cell interaction analysis 42 comparing TIME ligands and tumor cell receptors in EED -Mutant LUADs versus EED -WT SCLCs ( Fig. 4P-Q ), revealing 58 enriched receptor–ligand pairs. Cross-referencing this list with RNA-seq data from EED-inactivated 1014 cells ( Fig. 3L ) identified 14 upregulated tumor-expressed receptors with corresponding TIME ligands, including LUAD RTK oncogenes such as Met , Ret , and Fgfr2 ( Fig. 4P-Q ), all of which converge on RAS signaling. Notably, Met was the second-highest expressed receptor, with its canonical ligand Hgf highly expressed by macrophages in EED -Mutant RPR2 LUADs. These findings identify NEUROD1-positive tumor cells as a transitional state in the lineage switch from SCLC to LUAD following EED loss, and demonstrate that the tumor immune microenvironment promotes LUAD RTK oncogenic signaling to complete this transformation in vivo ( Fig. 4R ). EED Directly Binds and Represses RAS, PI3K, and MAPK Pathway LUAD Oncogenic Signaling Genes To elucidate how EED maintains the ASCL1-positive SCLC neuroendocrine cell state and why its loss drives SCLC-to-LUAD transformation, we performed ChIP-seq for EED and H3K27me3, identifying 6,072 direct EED target genes ( Fig. 3N ). Of these, 831 were epigenetically activated and expressed in EED -Mutant RPR2 LUADs ( Fig. 5A ), including LUAD oncogenes ( Met ) and gastrointestinal lineage genes ( Hnf4a , Cdx2 ) ( Fig. 5B ). Gene signatures enriched among these activated targets included RAS, PI3K, and MAPK signaling, EMT, and gastric cancer features ( Fig. 5C–D ), with strong overlap in inflammatory subtypes of human SCLC ( Fig. 5E ). Download figure Open in new tab Figure 5. EED Directly Represses Bivalently Marked RAS, PI3K, and MAPK Pathway Genes to Suppress LUAD Oncogenic Signaling and Enforce SCLC Neuroendocrine Identity. A . Schematic showing strategy to integrate direct EED target genes with genes reactivated upon EED loss using RNA-seq, ATAC-seq and H3K27ac ChIP-seq to identify direct functional EED targets in EED -Mutant LUAD lung tumors. B . Rank plots of RNA-seq log2FoldChange of functional EED targets in EED -Mutant LUAD lung tumors. C,D. Dot plots of KEGG pathway enrichment analysis ( C ) and hallmark pathway enrichment analysis ( D ) of direct functional EED targets in EED -Mutant LUAD. EMT: Epithelial to mesenchymal transition. BH adjusted p values were visualized by the dot color, red=low, blue=high. Enriched gene counts were visualized by the size of the dots. E . Gene set enrichment score of direct functional EED targets in EED -Mutant LUAD on different subtypes of human SCLC patient samples RNA-seq from IMPOWER133 (Nabet et al., 2024 12 ). SCLC-N=NEUROD1+ SCLC, SCLC-A=ASCL1+ SCLC, SCLC-I-NE=Inflammatory Neuroendocrine SCLC, SCLC-I-non-NE=Inflammatory non-Neuroendocrine SCLC. One way ANOVA test with multiple comparisons was used. *=p<0.05, **=p<0.01, ***=p<0.001, ****=p<0.0001. F-H . Schematic ( F ), KEGG pathway enrichment analysis ( G ), and gene network analysis of enriched Ras, AKT, and MAPK pathways ( H ) of overlapping EED direct functional targets in EED -Mutant SCLC and EED -Mutant LUAD. Enriched gene counts were visualized by the size of the dots. BH adjusted p values were visualized by the bar color, red=low, blue=high. I . Tracks of averaged RNA-seq, ATAC-seq, and H3K27ac ChIP-seq data across all samples from the EED -WT SCLC (blue), EED -Mutant SCLC (yellow) and EED -Mutant LUAD (red) mouse lung tumors and EED (brown) and H3K27me3 (purple) ChIP-seq data from EED -WT SCLC at Rasgrp1. J. Schematic showing genes with bivalent H3K4me3 and H3K27me3 histone marks and primed for reactivation upon PRC2 loss. Figure was made with BioRender. K,L . Schematic showing strategy to identify bivalent genes reactivated upon EED loss by overlapping peaks of H3K4me3 and H3K27me3 ChIP-seq data from EED -WT RPR2 SCLC tumors ( K ) and then overlap these bivalent peaks with reactivated EED direct target genes in EED -Mutant RPR2 LUADs ( L ). M . Expression of reactivated EED direct targets that are bivalent or not in EED -Mutant LUAD. N . Bar plot of enriched KEGG pathways of reactivated bivalent EED direct targets in EED -Mutant LUAD. BH adjusted p-values are visualized by the bar color, red=low, blue=high. Count: Enriched gene counts. O-Q . Tracks of averaged H3K27me3 (red) and H3K4me3 (green) ChIP-seq and RNA-seq at Akt1 (left), Mapkapk3 (middle), and Erbb2 (right) showing bivalent peaks and RNA-seq activation upon EED loss. See also Figure S6. Only 151 EED target genes were commonly activated in both EED -Mutant RPR2 LUADs and SCLCs ( Fig. 5F ), yet this shared set was significantly enriched for RAS and PI3K-AKT pathway genes ( Fig. 5G-H ). For instance, RASGRP1 , a RAS guanine nucleotide exchange factor, was a direct EED-repressed gene epigenetically activated in both contexts ( Fig. 5I ). These data suggest that EED directly represses RAS pathway components, and its loss upregulates RAS signaling even in tumors retaining SCLC histology, potentially priming for LUAD transformation. RAS pathway activation is a conserved hallmark downstream of nearly all LUAD oncogenic RTKs and has been shown to antagonize the NE phenotype in murine SCLC 43 —a finding we validated in human ASCL1-positive SCLC cell lines overexpressing wild-type or mutant (G12C) KRAS ( Fig. S6A–B ). To investigate whether EED loss induces RAS pathway dependence, we derived five cell lines from EED -Mutant RPR2 LUAD tumors (1350, 1339, 1343, 1344, 1345), which, unlike SCLCs, grew as adherent monolayers ( Fig. S6C–D ). These lines exhibited elevated phospho-ERK levels and showed strong sensitivity to the pan-RAS inhibitor RMC-6236 44 and the MEK inhibitor Trametinib, compared to EED -WT RPR2 SCLC cells ( Fig. S6E–H ), consistent with high RAS signaling and dependency. To test whether this phenotype was reversible, we re-expressed wild-type EED or a catalytically inactive mutant (EED-inactive) in two EED- Mutant LUAD lines (1339 and 1344) and performed RNA-seq ( Fig. S7A–B, S7E ). Re-expression of EED -WT, but not EED -inactive, downregulated PRC2 (EZH2/SUZ12) target genes, validating on-target activity ( Fig. S7C, S7F ). KEGG enrichment analysis further confirmed suppression of MAPK and RAP1 signaling pathways upon EED -WT re-expression ( Fig. S7D, S7G ). Together, these data demonstrate that EED directly represses RAS signaling to maintain the SCLC neuroendocrine phenotype, and that its loss induces RAS activation and LUAD oncogenic signaling, facilitating histological transformation. Loss of the PRC2 Complex Initiates SCLC-to-LUAD Histological Transformation Through Epigenetic Activation of Bivalent Genes Enriched in PI3K and MAPK Signaling While 6,072 genes were marked by both EED and H3K27me3, relatively few were activated upon EED loss—289 in SCLC and 831 in LUAD ( Fig. 3O , 5A ). Bivalent genes, which carry both the activating H3K4me3 and the repressive H3K27me3 histone marks, are known to be selectively de-repressed upon PRC2 inhibition during differentiation 45 . We therefore hypothesized that bivalent genes would be preferentially activated upon EED loss and include key regulators of SCLC-to-LUAD histological transformation ( Fig. 5J ). To identify bivalent genes, we performed ChIP-seq for H3K4me3 in EED -WT RPR2 SCLC tumors and overlaid these peaks with H3K27me3 peaks in the same samples, identifying 7,622 gene promoters with bivalent chromatin marks ( Fig. 5K , S8A–B ). As in previous studies 46 , 47 , GREAT and hallmark enrichment analyses showed that bivalent genes were associated with MHC class I antigen presentation and EMT ( Fig. S8C–E ). In line with our hypothesis, a large percent of genes activated upon EED loss had bivalent marks at their promoters including 118 of 289 (41%) genes activated upon EED loss in SCLC and 417 of 831 (∼50%) genes activated upon EED loss in LUAD ( Fig. 5L , S8F ). Consistent with genes marked by bivalency being poised for activation upon H3K27me3 loss 45 , reactivated EED direct targets in both EED -Mutant SCLC and LUAD that are bivalent have significantly higher expression than those that are not bivalent ( Fig. 5M , S8G ). Strikingly, bivalent genes activated in LUADs following EED loss were significantly enriched in PI3K-Akt and MAPK signaling pathways including Akt1 , Mapkapk3 , as well as upstream growth factor receptors (e.g. Erbb2 ) ( Fig. 5M-Q ). These findings indicate that EED directly represses genes that promote RAS pathway activation and that these genes are bivalent and poised for activation—thereby rapidly initiating SCLC-to-LUAD histological transformation upon PRC2 inactivation. On the other hand, Neurod1 is a bivalent target reactivated in EED -Mutant RPR 2 SCLC and among the top highly expressed genes ( Fig. S8F-H ). Some reactivated EED direct targets were not bivalent including the gastrointestinal lineage genes Hnf4A and Cdx1 ( Fig. S8F-H ) suggesting epigenetic activation of these genes requires further selection in vivo. This is consistent with our RNA-seq data in 1014 cells as Hnf4 and Cdx1 were not upregulated upon EED inactivation in vitro ( Fig. 3L ). Taken together, these findings suggest that bivalency marked by H3K27me3 and H3K4me3 poises RAS/PI3K/MAPK pathway genes and Neurod1 for activation, all of which are robustly induced upon EED loss. EED Inactivation Blocks LUAD to SCLC Transformation in an EGFR -Mutant GEMM as a Mechanism of Resistance to EGFR Inactivation Our earlier data showed that EED is required to maintain ASCL1-positive SCLC histology in the RPR2 GEMM, with EED inactivation promoting lineage infidelity and histological transformation from SCLC to LUAD ( Fig. 1 ). Based on this, we hypothesized that EED/PRC2 activity may also be required for the acquisition of SCLC histology during LUAD-to-SCLC transformation in the setting of resistance to EGFR inhibition, a process observed in EGFR -Mutant LUADs harboring RB1 and TP53 mutations 14 . To test this, we developed an EGFR -Mutant LUAD GEMM capable of histological transformation. We used a doxycycline (Dox)-inducible EGFR L858R transgenic model 48 , 49 crossed with: 1) Trp53 (P) flox/flox mice; 2) lox-stop-lox (LSL) rtTA mice to activate the Dox inducible promoter; and 3) LSL-Cas9-P2A-GFP mice ( Fig. 6A ). We then generated adenoviruses encoding: 1) Cre to activate Cas9 and rtTA expression; 2) sgRNAs targeting Rb1 (R). We also included an sgRNA targeting either EED (E) or a non-targeting sgRNA as a control (C) to ultimately make EGFR -Mutant LUAD GEMMs with concurrent Rb1 and Trp53 inactivation with EED inactivated (referred to hereafter as EGFR -Mutant ERP ) or where EED was WT (referred to hereafter as EGFR -Mutant CRP ). Immunoblot analysis validated that these adenoviruses induced Cre expression and efficiently inactivated RB1 and EED ( Fig. S9A ). Download figure Open in new tab Figure 6. EED Inactivation Blocks LUAD to SCLC Transformation as a Mechanism of Resistance to EGFR Inactivation in an EGFR -Mutant GEMM. A. Schematic of the CRISPR-based EGFR -Mutant LUAD GEMM with concurrent Rb1 and Trp53 inactivation made isogenic to Eed by including an sgT to target Eed to make ERP tumors or by including a non-targeting sgRNA as a control to make CRP tumors. Dox=doxycycline. LSL=lox-stop-lox. B . Schematic of adenovirus injection and Dox induction and withdrawal timeline. C . Kaplan-Meier survival estimate of mice IT injected with adenoviruses indicated and then treatment as indicated in (B). Log-rank p-value was used to calculate p-values. D . Bar plot of percent of mice with metastases present in each group. n=14 EGFR ON CRP independent mice, 12 EGFR OFF CRP independent mice, 14 EGFR ON ERP independent mice, and 16 EGFR OFF ERP independent mice. E-F . Pie charts of histology ( E ) with representative H&E and IHC staining for GFP, EGFPL858R, EED, ASCL1, NEUROD1 and SYP ( F ) of lung tumors from the groups indicated. Scale bar (1 st column) =1 mm. Scale bar (2 nd -8 th columns) =100 microns. Insets are 9X magnification. G . Heatmaps of z-scored gene expression of SCLC NE and non-NE markers from RNA-seq data of the EGFR -Mutant lung tumor genotypes indicated. Each row is an individual tumor. Black is low, yellow is high. n=3 for EGFR ON CRP lung tumors, n=9 for EGFR OFF CRP lung tumors, n=2 for EGFR ON ERP lung tumors, n=3 for EGFR OFF ERP lung tumors. H. Heatmaps of gene set enrichment scores of the indicated gene lists in EGFR -Mutant lung tumor genotypes indicated. Blue is low, yellow is high. IMA=invasive mucinous adenocarcinoma. TAM=tumor associated macrophages. I . Gene set enrichment score of MAPK target genes (Wagle, et al., 2018 67 ) in the EGFR -Mutant lung tumor genotypes indicated. Student’s t test was used to calculate two-sided p-values. J . PCA plot of RNA-seq data from the EGFR -Mutant lung tumor genotypes indicated. CRP: sgControl; sg Rb1 ; Trp53 -/-, ERP: sg Eed ; sg Rb1 ; Trp53 -/-. *=p<0.05, **=p<0.01, ***=p<0.001, ****=p<0.0001. EGFR ON=Dox ON, EGFR OFF=Dox withdrawal. See also Figure S9. These adenoviruses were intratracheally (IT) injected into the mice above and 1 week following IT injection, all mice were maintained on Dox-containing chow to induce expression of the EGFR L858R oncogene. Pilot experiments showed that after 20 weeks on Dox-containing chow, mice developed symptoms of dyspnea related to lung tumors with lung pathology showing EGFR -Mutant LUAD ( Fig. S9B-C ), which is consistent with the latency observed in this EGFR L858R transgenic mouse model 48 , 49 . Therefore after ∼20 weeks, both EGFR -Mutant CRP and EGFR -Mutant ERP mice were randomized to two groups with one group being maintained on Dox (Dox ON) and the other group switched back to regular chow (Dox OFF) to turn off expression of the EGFR L858R oncogene. Animals were then followed for symptoms until they reached their endpoint at which point necropsies were performed ( Fig. 6B ). Consistent with EGFR L858R acting as a potent oncogene, both EGFR -Mutant CRP and EGFR -Mutant ERP Dox ON mice more rapidly succumbed to their disease with a median overall survival of 30.3 weeks for EGFR -Mutant CRP and 31.0 weeks for EGFR -Mutant ERP mice ( Fig. 6C ). Survival was significantly prolonged in both EGFR -Mutant CRP and EGFR -Mutant ERP Dox OFF mice with a median overall survival of 57.9 and 57.1 weeks, respectively ( Fig. 6C ). Interestingly, the phenotypes of EGFR -Mutant CRP Dox OFF and EGFR -Mutant ERP Dox OFF mice were quite distinct. Tumors eventually recurred in EGFR -Mutant CRP Dox OFF mice with a significant number of mice (7 of 12) developing spontaneous distant metastases to the liver, kidney, and brain, while none of the EGFR -Mutant ERP Dox OFF mice developed spontaneously metastases ( Fig. 6D , S9D ). This is consistent with our observation that EED inactivation in the SCLC RPR2 GEMM blocked distant metastasis ( Fig. 1F ). IHC validated expression of the EGFR L858R transgene according to Dox treatment, GFP expression in tumor cells, and loss of EED protein only in EGFR -Mutant ERP mice but not in EGFR -Mutant CRP mice ( Fig. S9E ). DNA sequencing and immunoblot analysis confirmed expect patterns of inactivation in EGFR -Mutant CRP and EGFR -Mutant ERP tumors ( Fig. S9C,F,G ). Consistent with concurrent RB1 and TP53 mutations significantly increasing risk for SCLC transformation in human EGFR -Mutant LUAD 14 , some tumors in the EGFR -Mutant CRP Dox OFF mice (3 of 12) recurred as SCLC consistent with SCLC transformation where no tumors in the EGFR -Mutant ERP Dox OFF mice recurred as SCLC ( Fig. 6E ) suggesting that EED inactivation blocks LUAD-to-SCLC histological transformation. IHC demonstrated that transformed SCLC tumors in the EGFR -Mutant CRP Dox OFF mice either expressed only ASCL1 or heterogeneously expressed ASCL1 and NEUROD1, while the EGFR -Mutant ERP counterparts were completely negative ( Fig. 6F ), which is consistent with EGFR -Mutant human transformed SCLCs showing heterogenous expression of transcription factors with enrichment of variant subtypes including NEUROD1 50 . This also provided evidence to support that these tumors transformed into SCLC as a mechanism of escape from loss of EGFR L858R rather than formed as de novo SCLC as NEUROD1 expression is not observed in the RPR2 SCLC GEMM 20 , 21 , 28 , 51 . Consistent with the EED -Mutant RPR2 GEMM model, both EGFR -Mutant ERP Dox ON and Dox OFF tumors also showed evidence of LUAD with mucinous features ( Fig. 6F ). In line with this, there were only a small shared subset of 36 genes marked by bivalency in both human SCLC and human LUAD and this shared subset were highly enriched in drivers of gastric differentiation including FGFR2 and FGF18 ( Fig. S9H ). Consistent with our IHC data, bulk RNA-seq data showed increased expression of NE markers including Syp, Insm1 , and Ncam1 as well as Ascl1 with heterogenous Neurod1 and enrichment for SCLC-related signatures (e.g. MYC, G2M pathways, pancreas beta cell signatures) in transformed SCLCs relative to EGFR -Mutant CRP Dox ON tumors, EGFR -Mutant CRP Dox OFF tumors that did not transform to SCLC, or all EGFR -Mutant ERP tumors ( Fig. 6G-H , S9I-K ). Moreover, transformed SCLCs had lower MAPK target gene expression ( Fig. 6I ). EGFR -Mutant CRP Dox OFF LUADs were higher grade tumors relative to EGFR -Mutant CRP Dox ON LUADs and enriched expression of genes involved in EMT consistent with its poorly differentiated histology and higher propensity to metastasize ( Fig. S9L, see also Fig. 6D ). In line with findings in EED -Mutant RPR2 tumors, genes expressed in mucinous LUAD, tumor associated macrophages, and neutrophils were significantly enriched in both EGFR -Mutant ERP Dox ON and OFF tumors relative to their EGFR -Mutant CRP counterparts ( Fig. 6H , S9M-N ). Together, these data demonstrate that EGFR -mutant LUADs with RB1 and TP53 loss can evade EGFR dependency through either SCLC transformation or poorly differentiated metastatic LUAD. EED inactivation blocks both escape routes—preventing SCLC transformation and aggressive LUAD recurrence—and instead promotes a mucinous LUAD phenotype as a distinct mechanism of resistance to EGFR oncogene withdrawal ( Fig. 6J ). PRC2 Represses LUAD Oncogenic Programs in Human SCLC Through Bivalent Silencing of RAS, PI3K, and MAPK Pathway Genes To evaluate the clinical relevance of our findings and determine whether PRC2-mediated repression of LUAD oncogenic programs is conserved in human tumors, we analyzed H3K27me3 and H3K4me3 ChIP-seq data from 5 patient-derived xenograft (PDX) models of de novo neuroendocrine (NE) SCLC and 4 PDXs of de novo EGFR -Mutant LUAD 52 ( Fig. 7A-B , 7F) . EGFR -Mutant LUAD was selected given its established susceptibility to SCLC transformation 3 , making it a pertinent comparator for lineage plasticity. This allowed us to test whether PRC2 activity differentially regulates lineage-specific epigenetic programs in human lung cancer. Download figure Open in new tab Figure 7. PRC2 Represses LUAD Oncogenic Programs in Human SCLC Through Bivalent Silencing of RAS, PI3K, and MAPK Pathway Genes. A . Schematic showing gain of PRC2-mediated repression of genes with bivalent H3K4me3 and H3K27me3 histone marks. Figure was made with BioRender. B, F . PCA plot of H3K27me3 ChIP-seq ( B ) and H3K4me3 ChIP-seq ( F ) from 5 independent human de novo SCLC neuroendocrine patient-derived xenograft (PDX) tumors and 4 independent human EGFR -Mutant LUAD PDXs tumors. C,G . Log ratio/average (MA) plot comparing EGFR -Mutant LUAD PDX tumors vs. de novo SCLC neuroendocrine (NE) PDX tumors above to visualize the relationship between the log ratio (M) and the average (A) of H3K27me3 ChIP-seq data ( C ) and H3K4me3 ChIP-seq data ( G ). D,H . Box plots of normalized read counts of H3K27me3 ChIP-seq ( D ) and H3K4me3 ChIP-seq ( H ) in the indicated groups. E . Bar plot of enriched KEGG pathways in genes mapped to the differential H3K27me3 peaks gained in NE SCLC PDX tumors vs. EGFR -Mutant LUAD PDX tumors. BH adjusted p-values are visualized by the bar color, red=low, blue=high. Count: Enriched gene counts. I . Bar plot of ENCODE and ChEA Consensus TFs enrichment score from top enriched transcription factors whose targets were enriched in genes mapped to the differential H3K4me3 peaks that gained signal in NE SCLC PDX tumors. Yellow: significantly enriched with BH adjusted p-value<0.05. J . Schematic to identify bivalent genes marked by both H3K4me3 and H3K27me3 in human NE SCLC PDX tumors. K . Bar plot of enriched KEGG pathways in genes marked with bivalent H3K4me3 and H3K27me3 in human NE SCLC PDX tumors. BH adjusted p-values were visualized by the bar color, red=low, blue=high. Count: Enriched gene counts. L . Gene network analysis of enriched RAS, PI3K-AKT, and MAPK pathways from P . Enriched gene counts are visualized by dot size. M . Bar plots indicating the percentage of genes marked by H3K4me3 of the indicated pathways that gained H3K27me3 in NE SCLC PDX tumors and and hence are bivalent. N . Density plots of normalized read count of H3K4me3 (top) and H3K27me3 (bottom) ChIP-seq peaks mapped to genes that belong to KRAS signaling UP, PI3K-Akt and MAPK KEGG pathways in human NE SCLC and EGFR -mutant LUAD PDX tumors. O-Q . Bar plots of RNA-seq batch normalized expression of EED ( O ), EZH2 ( P ), SUZ12 ( Q ) from human EGFR -Mutant LUAD tumors with the RB1 and TP53 genetic alterations indicated from the Cancer Genome Atlas (TCGA) 68 . RB1/TP53 WT: n=22 tumors, TP53 mutation or deep deletion: n=34 tumors, RB1/TP53 mutation or deep deletion: n=10 tumors. One way ANOVA with Tukey test was used to calculate two-sided p values adjusted for multiple comparisons. *=p<0.05, **=p<0.01, ***=p<0.001, ****=p<0.0001. See also Figure S10. ChEA and ENCODE transcription factor analyses of H3K27me3 peaks identified SUZ12 and EZH2 as dominant regulators of both histologies ( Fig. S10A-B ), validating the fidelity of the ChIP-seq dataset. Consistent with our data from mouse tumors and previous studies 46 , 47 , GO analysis showed MHC I antigen presentation genes were enriched for H3K27me3 in SCLC, while axon guidance pathways were enriched for H3K27me3 in LUAD ( Fig. S10C-D ) together showing that H3K27me3 gene repression patterns helps define key differences between SCLC and LUAD. De novo NE SCLC had overall more and stronger H3K27me3 peaks relative to LUAD and the top gained H3K27me3 peaks in SCLC vs. LUAD included PI3K and MAPK pathway genes ( Fig. 7C-E )—mirroring findings from our GEMM studies. SCLC also showed elevated H3K4me3 at neurodevelopmental genes ( Fig. 7G–H , S10E-F ), whereas LUAD-specific H3K4me3 peaks enriched for EMT, inflammatory response, and KRAS signaling ( Fig. S10G-I ). Unexpectedly, the strongest H3K4me3 peaks gained in SCLC enriched for PRC2 targets, suggesting that increased PRC2 activity may coincide with H3K4me3 deposition at poised, bivalent loci ( Fig. 7I ). To test this, we identified bivalent promoters (co-marked by H3K27me3 and H3K4me3) in human SCLC and LUAD. We found 548 bivalent genes in SCLC and only 202 in LUAD ( Fig. 7J , S10J ). Bivalent genes in SCLC were significantly enriched for RAS, PI3K-AKT, and MAPK signaling, including RASGRP1 ( Fig. 7K–L ), and were largely distinct from those in LUAD ( Fig. S10K-L ). Notably, ∼90% of all RAS/ PI3K-AKT/MAPK pathway genes that gained H3K27me3 in SCLC were bivalent ( Fig. 7M ), and genome-wide H3K27me3 signal was substantially higher across these pathways in SCLC than LUAD ( Fig. 7N ). Finally, to investigate why PRC2 activity is elevated in SCLC, we analyzed expression of PRC2 components in human LUAD samples from TCGA. Given that RB1 and TP53 loss is nearly universal in SCLC 11 and common in LUADs that undergo transformation to SCLC 14 , and that RB1 loss promotes E2F transcription factor activity, which directly drives expression of PRC2 components such as EZH2 and EED 53 , we hypothesized that RB1 inactivation is associated with elevated PRC2 expression. Indeed, RB1/TP53 -mutant LUADs showed significantly higher expression of all three core PRC2 components (EZH2, EED, and SUZ12) compared to RB1/TP53 -WT LUAD tumors ( Figs. 7O-Q , S10M-O ). Together, these data provide human tumor evidence that PRC2 blocks LUAD oncogenic potential in SCLC by repressing key signaling effectors through bivalent chromatin, and suggest a mechanistic axis in which RB1 and TP53 loss drives PRC2 overexpression to silence bivalent RAS, PI3K, and MAPK genes—thereby enforcing the neuroendocrine phenotype. Discussion Here, we demonstrate that EED, a core component of the PRC2 complex, is essential for maintaining the ASCL1-positive NE identity and histological state of SCLC during tumorigenesis. When tumorigenesis is initiated in the RPR2 genomic context that normally drives formation of SCLC tumors, tumors escape from EED inactivation by losing ASCL1 and adopting LUAD histology. Alternatively, in an EGFR -Mutant LUAD GEMM with concurrent Rb1 and Trp53 loss—genetic lesions strongly associated with LUAD-to-SCLC transformation 14 —EED was required for SCLC histology to emerge following EGFR oncogene withdrawal. Together, these findings identify EED as a critical determinant of the SCLC NE histological fate in lung cancer. Mechanistically, EED directly binds and promotes H3K27me3-mediated gene silencing of RAS, PI3K, and MAPK signaling genes to repress LUAD oncogenic signaling and maintain SCLC NE identity. These loci are bivalent—marked by both H3K27me3 and H3K4me3—in SCLC, rendering them epigenetically poised for rapid activation. Loss of EED triggers SCLC-to-LUAD histological transformation by derepressing these bivalent targets, unleashing LUAD oncogenic signaling through a NEUROD1-positive intermediate state that requires cues from the tumor microenvironment ( Fig. 8 ). Supporting this model, ChIP-seq analysis of human NE SCLC PDXs revealed bivalent chromatin marks at the same RAS, PI3K, and MAPK pathway genes, indicating dynamic PRC2-mediated repression in human disease. This supports a model in which PRC2 activity silences LUAD oncogenic signaling to enforce and maintain SCLC identity. Download figure Open in new tab Figure 8. Schematic Showing How EED Constrains the SCLC Neuroendocrine Phenotype and Drives Lung Cancer Histological Transformation. When SCLC tumorigenesis is initiated in the RPR2 GEMM model, inactivation of EED delays tumor formation and tumors eventually form by a histological transformation to lung adenocarcinoma (LUAD). SCLC-to-LUAD histological transformation occurs through a transient NEUROD1-positive EMT-like intermediate cell state that requires cues from the tumor microenvironment. When LUAD tumors are initiated by the EGFR L858R oncogene and recur after EGFR oncogene withdrawal, only tumors with EED-intact undergo histological transformation to SCLC as a mechanism of resistance to EGFR withdrawal. Together both models show that intact EED is strictly required to promote the SCLC neuroendocrine phenotype during lung cancer pathogenesis. Mechanistically, EED maintains the SCLC neuroendocrine identity by directly repressing bivalent chromatin-marked RAS, PI3K, and MAPK pathway genes to oppose LUAD oncogenic signaling. RTK: receptor tyrosine kinase. EMT: epithelial to mesenchymal transition. Figure created with BioRender. Human lung cancers can exist as combined LUAD and SCLC histology tumors and can also histologically transform from LUAD to SCLC as a mechanism of resistance to targeted therapy 2 , 3 . Our data suggests that mechanisms to promote PRC2 complex activity would select for SCLC histology in each of these settings. Although SCLC-to-LUAD transformation has not been observed in human tumors, this may reflect the limited efficacy of current SCLC therapies, which exert less selective pressure to escape the SCLC neuroendocrine state. Recently the highly effective neuroendocrine targeted therapy Tarlatamab, a DLL3 bispecific T cell engager, was approved 54 , 55 . It will be important to determine whether resistance to Tarlatamab could involve SCLC-to-LUAD histological transformation as a strategy to escape the neuroendocrine state and promote tumor survival. EED inactivation in both SCLC RPR2 GEMMs and EGFR -Mutant RP LUAD GEMMs led to LUAD with aberrant expression of gastrointestinal differentiation genes. Similarly, prior work showed that EED loss promoted mucinous LUAD histology in KRAS -mutant LUAD GEMMs 56 , 57 , suggesting that PRC2 normally represses GI differentiation programs in lung tumors. We speculate that this function of PRC2 mirrors the foregut endoderm anterior to posterior patterning where it likely represses genes involved in GI differentiation such as HNF4A , WNT and FGF signaling to give rise to the respiratory system and inhibit gut development 58 . We found that EED inactivation also de-repressed NEUROD1 in vitro and occasionally drove ASCL1-to-NEUROD1 subtype switching in SCLC tumors in vivo . NEUROD1 is a direct EED target marked by bivalent H3K4me3 and H3K27me3, enabling rapid activation upon PRC2 loss. While ASCL1 is expressed in pulmonary neuroendocrine cells 59 , NEUROD1 is absent in the lung but highly expressed in endocrine pancreatic β cells 60 – 62 , which are also derived from foregut endoderm 63 . There are several studies that now show ASCL1 to NEUROD1 subtype switching in SCLC and NEPC after selective pressure against ASCL1 20 , 21 , 64 . Our findings reveal another mechanism that drives ASCL1 to NEUROD1 subtype switching and suggests that this can occur when tumor cells adopt development programs expressed in pancreatic β cells. While our previous work showed that the epigenetic modifier KDM6A maintains an epigenetic state permissive for the ASCL1 subtype by preferentially through epigenetic activation of genes associated with the ASCL1 subtype 21 , this works shows that the PRC2 complex achieves an ASCL1-positive subtype by repressing genes associated with the NEUROD1 subtype. Following EED inactivation, NEUROD1-positive SCLC tumors were rarely sustained, and nearly all mice ultimately developed LUADs, suggesting that NEUROD1-positive SCLCs are either transient and selected against, or that LUAD histology is actively selected for. We observed NEUROD1-positive cells within LUADs that emerged after EED loss, and our snRNA-seq data identified a NEUROD1-expressing, EMT-like subpopulation, supporting a model in which LUADs arise from SCLC through a NEUROD1-positive intermediate. Notably, EED inactivation alone was insufficient to drive this transition in vitro , implying that the tumor immune microenvironment (TIME) is required to complete SCLC-to-LUAD histological transformation, potentially by enhancing LUAD oncogenic signaling. Alternatively, prior work, including our own, has shown that PRC2 loss in SCLC increases MHC class I expression 46 , 47 , which correlates with enhanced immune visibility and response to checkpoint blockade 46 , 65 . It is therefore possible that EED-inactivation renders SCLC cells immunogenic and subject to immune-mediated clearance, thereby favoring outgrowth of LUAD. Consistent with this, EED inactivated tumors exhibited a marked increase in immunosuppressive M2-like macrophages, which may emerge in response to restored MHC class I expression and facilitate immune evasion. Future studies will aim to dissect how the TIME contributes to lineage plasticity in lung cancer and whether transformation between lung cancer histological subtypes reflects tumor-intrinsic reprogramming, immune selection, or both. LUAD-to-SCLC transformation is a well-established mechanism of resistance to targeted therapies, most frequently observed following EGFR inhibition in EGFR -Mutant LUADs with co-occurring RB1 and TP53 mutations 2 , 3 , 5 . While these alterations increase the relative risk of transformation by 43-fold 14 , only ∼18% of such cases undergo SCLC transformation clinically 8 . In our study, EED inactivation fully blocked the ability of EGFR -Mutant LUADs with Rb1 and Trp53 loss to transform into SCLC following EGFR oncogene withdrawal. Although only ∼25% of tumors in our model exhibited SCLC histology—closely aligning with the rate in EGFR -Mutant RB1 / TP53 mutated human tumors—most remaining tumors recurred as poorly differentiated, metastatic LUADs, highlighting divergent lineage escape routes. This incomplete penetrance likely reflects the absence of additional co-occurring events, such as PI3K pathway activation 5 , 10 , 15 or MYC overexpression 10 , which are frequently found in human LUADs that undergo lineage transformation. Indeed, a recent EGFR -Mutant GEMM incorporating non-degradable c-MYC showed a higher frequency of LUAD-to-SCLC transformation 10 . Our study identifies PRC2 as a central epigenetic gatekeeper of lineage plasticity in lung cancer, maintaining the SCLC neuroendocrine phenotype by repressing LUAD oncogenic signaling through bivalent chromatin silencing of RAS, PI3K, and MAPK pathway genes. Integrating murine models with human PDX ChIP-seq data, we demonstrate that these signaling genes are marked by both H3K27me3 and H3K4me3 in SCLC, positioning them for rapid activation upon PRC2 loss. Importantly, we show that RB1 and TP53 co-mutation—common in both de novo SCLC and LUADs at risk of SCLC transformation—is associated with elevated expression of PRC2 components (EZH2, EED, SUZ12). These findings suggest a mechanistic link between genetic drivers and chromatin-based lineage restriction, and implicate PRC2 activity as a driver of LUAD-to-SCLC transformation under therapeutic pressure. As such, targeting PRC2 may represent a rational strategy to prevent histological transformation in high-risk EGFR -Mutant LUADs with RB1 and TP53 loss, and more broadly, to modulate lineage plasticity in lung cancer. Limitations of the study Our data suggests that inhibition of the PRC2 complex with small molecule EZH1/2 or EED inhibitors as a therapeutic approach to block SCLC histological transformation in patients at high-risk for SCLC transformation; particularly patients with EGFR -Mutant LUADs with co-occurring RB1 and TP53 inactivation 8 , 14 . However, in our experiments, EED was genetically inactivated at tumor initiation in mice, not in established EGFR oncogene-driven tumors. Our analysis of ChIP-seq data from human LUAD and SCLC PDXs uncovering a gain of H3K27me3 in SCLC to silence genes involved in RAS, PI3K, and MAPK signaling, and drive SCLC histology supports this therapeutic rationale in human tumors, which is consistent with a previously published independent data in human tumors showing a gain of PRC2 complex transcriptional signatures during LUAD-to-SCLC transformation 5 . Future studies will focus on determining whether clinical grade inhibitors that disrupt the PRC2 complex can block histological transformation in established tumors driven by EGFR or other LUAD oncogenic drivers such as ALK rearrangements or KRAS mutations, where SCLC histological transformation has also been observed; albeit less commonly 66 . Author Contributions Y.L.: Conceptualization, methodology, validation, formal analysis, investigation, data curation, supervision, writing. Y.N.L, H.C., G.R.D.O., T.E.Z.: Methodology, validation, formal analysis, investigation, data curation. M.V., Y.D., A.D., X.Q., S.K., R.L., R.B., S.L., W.L., M.V.O.: Investigation. K.P., J.B., M.F., H.L., S.S.: Resources, visualization, supervision, project administration. Y.C., H.J.: Methodology. M.G.O: Conceptualization, methodology, investigation, resources, data curation, writing, visualization, supervision, project administration, funding acquisition. Declaration of Interests M.G.O. reports grants from Novartis, Circle Pharma, Amgen, Auron Therapeutics, Eli Lilly, Takeda, and BMS; none of which are related to this work. K. P. reports grants from AstraZeneca, Roche/Genentech, Boehringer Ingelheim, and D2G Oncology; Personal fees from AstraZeneca and Revelio Therapeutics, Inc; Patent related to EGFR T790M mutation testing with royalties paid “from MSKCC/MolecularMD”; Co-founder of and consultant for Revelio Therapeutics, Inc. M.F. is a Co-Founder and reports personal fees and other support from Precede Biosciences outside the submitted work. No other authors report relevant conflicts of interest. Methods All experiments herein comply with all ethical regulations. Specifically, all mouse experiments complied with National Institutes of Health guidelines and were approved by Dana-Farber Cancer Institute Animal Care and Use Committee (DFCI, protocol 19-009). All adenoviral and lentiviral transduction experiments complied with the Biohazard Control Committee (DFCI, protocol 19-1133). Adenoviral sgRNA Expression Vector Cloning to Create RPR2 Genetically Engineered Mouse Model Effective sgRNAs targeting mouse Rb1 , Trp53 , and Rbl2 were first validated using lentiviral vectors as described previously 26 . Effective sgRNAs targeting mouse Eed were validated in mouse embryonic fibroblasts expressing Cas9. The cloning method for generation of adenoviral sgRNA expression vectors encoding CMV-Cre recombinase and sgRNAs targeting Rb1, Trp53, and Rbl2 and “T” sgRNA (in this case sgEed#1, sgEed#2, or a non-targeting sgRNA control, C0111 ) was also described previously 25 , 27 . Briefly, a pENTR223-CMV-Cre-U6-sgX-U6-sgRb1-U6-sgTrp53-U6-sgRbl2 where X is sgEed#1, sgEed#2, or sgC0111, was used in an LR recombination reaction to clone the 4 pENTR223-CMV-Cre-U6-sgX-U6-sgRb1-U6-sgTrp53-U6-sgRbl2 vectors described above into pAd-PL DEST (Invitrogen) according to the manufacturer’s instructions. The recombinants were transformed into HB101 cells and ampicillin-resistant colonies were screened by restriction digestion of miniprep DNA and subsequently validated by DNA sequencing. The following sgRNA oligos were used (including BsmBI sites): Rb1 mouse #11 sense (5’- CACCGCAACTAGAAAATGATACG-3’), Rb1 mouse #11 anti-sense (5’- AAACCGTATCATTTTCTAGTTGC-3’), Trp53 mouse #8 sense (5’- CACCGGTGTAATAGCTCCTGCATGG-3’), Trp53 mouse #8 anti-sense (5’- AAACCCATGCAGGAGCTATTACACC-3’), Rbl2 mouse #6 sense (5’- CACCGAGGAGGATGGCGACGCCG-3’), Rbl2 mouse #6 anti-sense (5’- AAACCGGCGTCGCCATCCTCCTC-3’), Eed mouse #1 sense (5’- CACCGACAAATACGCCAAATGCACC-3’), Eed mouse #1 anti-sense (5’- AAACGGTGCATTTGGCGTATTTGTC-3’), Eed mouse #2 sense (5’-CACCGTGCACCAGGAAGGAAAAGCT-3’), Eed mouse #2 anti-sense (5’-AAACAGCTTTTCCTTCCTGGTGCAC-3’), C0111 ( Non-targeting sgRNA, sgControl) sense (5’- CACCGGGAGGCTAAGCGTCGCAA-3’), C0111 (Non-targeting sgRNA, sgControl) anti-sense (5’- AAACTTGCGACGCTTAGCCTCCC-3’). Adenoviral sgRNA Expression Vector Cloning to Create EGFR -Mutant Genetically Engineered Mouse Model with Concurrent Rb1 and Trp53 Mutation The cloning method for generation of adenoviral sgRNA expression vectors encoding CMV-Cre recombinase and sgRNAs targeting Rb1 and “T” sgRNA (in this case sgEed#1, sgEed#2, or a non-targeting sgRNA control) is described below. The DNA fragment containing CMV-Cre-U6-sgX-U6-sgRb1 was generated by inverse polymerase chain reaction (PCR) using KOD Xtreme™ Hot Start DNA Polymerase (Sigma-Aldrich, Cat# 71975) with primers (sense: 5’-CACCCAGCTTTCTTGTACAAAGTTGGC-3’, and anti-sense: 5’-TTGTACAAGAAAGCTGGGTGTTGTATA-3’) from the pENTR223-CMV-Cre-U6-sgX-U6-sgRb1-U6-sgTrp53-U6-sgRbl2 template. Linear DNA fragment was annealed into a circular plasmid by Gibson Assembly (NEB, cat# E5510S) according to the manufacturer’s instructions. The plasmid was then transformed in HB101 cells, and isolated using the QIAprep Spin Miniprep Kit – Plasmid Purification Kit (Qiagen, 27106). Whole plasmid sequencing was performed to confirm the correct sequences (pENTR223-CMV-Cre-U6-sgX-U6-sgRb1 where sgX is sgEed#1, sgEed#2, or a non-targeting sgRNA control). Subsequently LR reactions were performed with pAd-PL DEST as described above and previously 25 , 27 to make pAd-PL CMV-Cre-U6-sgX-U6-sgRb1 where sgX is sgEed#1, sgEed#2, or a non-targeting sgRNA control. Adenovirus Production and Purification Adenoviral production and purification were performed as described previously 25 . 5 μg of the adenovirus vector (pAd/PL Invitrogen #V494-20) containing the desired sgRNA sequences and Cre recombinase expression cassette (see above) was digested with PacI (New England Biolabs) for 2 hours at 37°C according to the manufacturer’s instructions and column purified using Qiagen’s gel extraction kit. 1 μg of PacI-digested pAd/PL was transfected into 1.5 X 10 6 293AD cells plated on a 6 cm tissue-culture dish using Lipofectamine 2000. The following day, the media was exchanged, and subsequently exchanged every 48 hours thereafter. Once 293AD cells showed evidence of adenovirus production (determined by comet formation with lysis), the cells and supernatant were harvested, which were then subjected to 4 freeze-thaw cycles by alternating between an ethanol dry ice bath and 37°C. Cell debris was removed by centrifugation and the supernatant was collected, passed through a 0.45 μm filter, aliquoted, and frozen at -80°C until use. To generate high titer adenovirus for in vivo experiments, adenovirus was generated as described above. 50 μl of the adenovirus stock was added to each 10 cm tissue-culture dish of 293FT cells plated at 3 X 10 6 cells per dish (4 10 cm dishes in total for each purification). When 293FT cells showed evidence of adenovirus production, as determined by cell rounding and partial detachment (∼48-72 hours after addition of adenoviral stock), the cells were collected, and adenovirus was purified using Virabind Adenovirus Purification Kit (Cell Biolabs #VPK-5112). The purified adenovirus was titered using QuickTiter Adenovirus Quantitation Kit (Cell Biolabs #VPK-106) according to the manufacturer’s instructions. Intratracheal Injections Intratracheal injections were performed as described previously 69 . Briefly, mice were anesthetized with ketamine and xylazine and pedal reflexes were monitored to ensure adequate anesthesia. Mice were maintained on a heated stage at 37° C while anesthetized. Mice were hung on stage with their top incisors and intubated with a 22-gauge 1 inch catheter (ThermoFisher Scientific #1484120). Once intubated, adenovirus (4 X 10 8 VP/mouse) in a total volume of 75 μl (diluted in PBS) was added to the catheter and subsequently inhaled by the mice. Establishing Genetically-Engineered Mouse Models using CRISPR/Cas9 For the RPR2 GEMM, pure congenic Lox-stop-lox (LSL) Cas9 BL6J mice were purchased from Jackson Labs (Jackson No. 026175) and maintained as homozygous BL6J mice. Genotyping of Cas9 and GFP at the ROSA26 were confirmed for all mice on the study (Transnetyx). Above-mentioned RPR2 adenovirus was injected into these mice to generate EED -isogenic RPR2 tumors. For the EGFR -Mutant GEMM with concurrent Rb1 and Trp53 Mutation, TetO-EGFR L858R ; Trp53 flox/flox ; Rosa26 CAGs-LSL-rtTA3-IRES-mKate and TetO-EGFR L858R ; Trp53 flox/flox ; Rosa26 CAGs-LSL-Cas9-GFP were a kind gift from the laboratory of Dr. Katerina Politi 49 . These mice are mixed BL6/129/FVB background and maintained as homozygous strains. These two strains were bred together to generate TetO-EGFR L858R ; Trp53 flox/flox ; Rosa26 CAGs-LSL-rtTA3-IRES-mKate ; Rosa26 CAGs-LSL-Cas9-GFP mice used in this study (referred to hereafter as EGFR -Mutant GEMM experimental mice). Genotyping of EGFR L858R , Trp53 flox/flox , and Cas9, GFP , and rtTA3 at the ROSA26 locus were confirmed for all mice (Transnetyx). For experiments in Fig. 6 and S9, pAd-PL CMV-Cre-U6-sgEed#2-U6-sgRb1 or pAd-PL CMV-Cre-U6-sgControl-U6-sgRb1 adenoviruses were intratracheally injected as above into the EGFR -Mutant GEMM experimental mice and then all mice were fed doxycycline-containing chow (625 ppm, cat# 5AW9, ScottPharma Solutions Inc.). For the control de novo LUAD tumors in Fig. 4A , RPR2 adenovirus were used to match the adenoviruses used to make the SCLC RPR2 model and intratracheally injected into the EGFR -Mutant GEMM experimental mice followed by administration of doxycycline-containing chow (625 ppm) after 1 week. For all studies with the RPR2 GEMM and the EGFR -Mutant GEMM, both male and female mice were used at roughly equal numbers. Housing conditions for mice at the DFCI Vivarium include a 12 hour/12 hour day-night cycle where temperature is maintained at 72 F. Roughly equal numbers of male and female 3-4 month-old mice were intratracheally injected with 4 X 10 8 VP/mouse adenovirus. All mice were monitored at least twice a week and euthanized when they became symptomatic (primarily respiratory distress), moribund, or lost 15% of their total body weight. The maximal tumor size allowed by the Dana-Farber Cancer Institute Animal Care and Use Committee is 2 cm and the maximal tumor size was not exceeded in any of our studies. Upon euthanization, half of the lung tumor specimen was immediately flash frozen on dry ice for subsequent DNA, RNA, and protein analysis, while the other half and the rest of lung was fixed in 10% formalin for 24 hours and then stored in 70% ethanol prior to paraffin embedding. For some tumors, cell lines were generated (see method below). Livers, kidneys, and brains were harvested and fixed and embedded as above. Slides were made for Hematoxylin and Eosin (H&E) and immunohistochemistry (IHC). H&E slides were analyzed by a specialized rodent pathologist Dr. Roderick Bronson, or pulmonary pathologist Dr. Marina Vivero for diagnosis. Generation of Cell Lines from Mouse Tumors and Cell Culture sgControl (EED-WT) RPR2 (631, 1014) cell lines were generated from CRISPR-based SCLC GEMMs (see above) as described previously 25 . Once tumors developed, mice were euthanized with CO2 and their tumors were quickly extracted, washed in ice cold PBS, and minced several times using an ethanol sterilized razor blade. 3mL of collagenase/hyaluronidase (Stem cell biology #07912) diluted 1:10 in complete RPMI media containing [10% FBS, P/S, and HITES (10 nM hydrocortisone (Sigma Aldrich # H0135), Insulin-Transferrin-Selenium (Gemini #400-145), and 10 nM beta-estradiol (Sigma Aldrich# E2257), 100 U/mL of penicillin (P), and 100 µg/mL of streptomycin (S)], and 1mL dispase (Corning # 354235) was added to the tumor, and incubated at 37°C for 20-40 minutes with periodic pipetting ∼10 every minutes (until most of the tumor cells were in suspension). The cells were then collected, centrifuged at 200 x g for 5 minutes, resuspended in complete RPMI media (see above), filtered through a 70 μm cell strainer (BD #352350), centrifuged again at 1000 rpm for 5 minutes, resuspended in fresh RPMI HITES media and placed in ultra-low adherence tissue culture dishes (Corning #3471). Media was subsequently replaced every 3 days. Histopathology on the tumors confirmed SCLC for all cell lines generated. All SCLC cell line were grown in Ultra-Low Attachment flasks (Corning™ 3814CONV) or plates (Corning™ 3471) at 37°C in the presence of 5% CO2. Once established, all cell lines were validated using immunoblot analysis for Cas9 and the SCLC neuroendocrine markers ASCL1. sg Eed RPR2 (1339, 1343, 1344, 1345, 1350) cell lines were generated from EED -Mutant RPR2 LUAD tumors using Mouse Tumor Dissociation Kit (Miltenyi, cat# 130-096-730) following manufacturer’s instructions. All LUAD RPR2 cell line were grown in RPMI media containing [10% FBS, 100 U/mL of penicillin (P), and 100 µg/mL of streptomycin (S) (P/S)] on tissue culture dishes (Falcon™, cat# 353003) at 37°C in the presence of 5% CO2. Once established, all cell lines were validated by immunoblot analysis for expression of Cas9, loss of TP53 and RB1 expression, and loss of ASCL1 expression, and by FACS for GFP expression. Early passage cell lines were tested for Mycoplasma (Lonza #LT07-218) and were negative, and then were frozen using Bambanker’s freezing media (Bulldog Bio). Mouse embryonic fibroblasts expressing Cas9 used to validate the adenoviruses were described previously 26 and cultured in DMEM media with 10% FBS and P/S. Human Cell Lines NCI-H1876, NCI-H2081 and 293FT cells were originally obtained from American Type Culture Collection (ATCC). 293AD cells (AD-100) were obtained from Cell Biolabs. NCI-H1876 and NCI-H2081 cells were maintained in DMEM/F12 media 5% FBS, P/S, and HITES. 293T and 293AD were maintained in DMEM media with 10% FBS and P/S. Early passage cell lines were tested negative for Mycoplasma (Lonza #LT07-218), and then were frozen using Bambanker’s freezing media (Bulldog Bio). All experiments were performed with cell lines that were maintained in culture for <3 months at which time an early passage cell lines were thawed. No commonly misidentified cell lines were used in this study. Pharmacological Inhibitors The following chemicals (stored at -20°C or -80°C) were added to cell culture where indicated: MAK683 (Selleckchem #S8983, stock 10 mM in DMSO), Tazemetostat (Selleckchem #S7128, stock 10 mM in DMSO), Tulmimetostat (Selleckchem #E1497, stock 10 mM in DMSO), Valemetostat (Selleckchem #S8926, stock 10 mM in DMSO), Trametinib (Selleckchem #S2673, stock in 10mM in DMSO) or RMC-6236 (Selleckchem #E1597, stock 10mM in DMSO). sgRNA Cloning to Make Lentiviruses sgRNA sequences were designed using the Broad Institute sgRNA designer tool ( http://portals.broadinstitute.org/gpp/public/analysis-tools/sgrna-design ) and synthesized by IDT technologies. The sense and antisense oligonucleotides were mixed at equimolar ratios (0.25 nanomoles of each sense and antisense oligonucleotide) and annealed by heating to 100°C in annealing buffer (1X annealing buffer 100 mM NaCl, 10 mM Tris-HCl, pH 7.4) followed by slow cooling to 30°C for 3 hours. The annealed oligonucleotides were then diluted at 1:400 in 0.5X annealing buffer. For CRISPR/Cas9 knockout experiments in cells, the annealed oligos were ligated into LentiGuide Puro (Addgene #52963) for experiments in mouse SCLC cell lines. Ligations were done with T4 DNA ligase for 2 hours at 25°C. The ligation mixture was transformed into HB101 competent cells. Ampicillin-resistant colonies were screened by restriction digestion of miniprep DNAs and subsequently validated by DNA sequencing. The following sgRNA oligos were used for LentiGuide Puro vector for CRISPR knockout experiments: Eed mouse #1 sense (5’- CACCGACAAATACGCCAAATGCACC-3’), Eed mouse #1 anti-sense (5’- AAACGGTGCATTTGGCGTATTTGTC-3’), Eed mouse #2 sense (5’-CACCGTGCACCAGGAAGGAAAAGCT-3’), Eed mouse #2 anti-sense (5’-AAACAGCTTTTCCTTCCTGGTGCAC-3’), C0111 (Non-targeting sgRNA, sgControl) sense (5’- CACCGGGAGGCTAAGCGTCGCAA-3’), C0111 (Non-targeting sgRNA, sgControl) anti-sense (5’- AAACTTGCGACGCTTAGCCTCCC-3’). Lentivirus Production Lentiviruses were made by Lipofectamine 2000-based co-transfection of 293FT cells with the respective lentiviral expression vectors and the packaging plasmids psPAX2 (Addgene #12260) and pMD2.G (Addgene #12259) at a ratio of 4:3:1. Virus-containing supernatant was collected at 48 and 72h after transfection, pooled together (15 mL total per 10-cm tissue culture dish), passed through a 0.45-µm filter, aliquoted, and frozen at -80°C until use. Lentiviral Infection The cells were counted using a Vi-Cell XR Cell Counter (Beckman Coulter) and 2 X10 6 cells were resuspended in 1mL lentivirus with 8 μg/mL polybrene in individual wells of a 12 well plate. The plates were then centrifuged at 448 x g for 2h at 30° C. 16 hours later the virus was removed, and cells were grown for 72 hours before being placed under drug selection. Cells were selected in puromycin (0.5 μg/mL). Generation of SCLC Xenograft Model in BL6J and NCr Nude Mice The syngeneic RPR2 1014 BL6J model was described previously 27 . 1014 cells were first transduced with plentiguide-puro lentiviruses containing an sgRNA targeting EED ( Eed mouse #2 sense (5’-CACCGTGCACCAGGAAGGAAAAGCT-3’), Eed mouse #2 anti-sense (5’-AAACAGCTTTTCCTTCCTGGTGCAC-3’)) (EED-inactivated) or a non-targeting sgRNA as a control (EED-WT). Immunoblot analysis validated loss of EED protein with increased NEUROD1 expression (see Fig. 3K ). Then 8x10 6 RPR2 1014 with EED -WT or EED -Inactivated cells were washed 3 times in DPBS and then resuspended in DPBS with 33% Matrigel (Corning cat# 354234) and implanted subcutaneously into 7-9 week-old female albino B6 mice (C57BL/6J-Tyrc-2J, stock no. 000058; RRID:IMSR_JAX:000058) from The Jackson Laboratory (Bar Harbor, ME) or immunodeficient female NCr nude mice (Taconic #NCRNU). Once established, mice were euthanized with CO2, and half of each tumor was immediately flash frozen on dry ice for subsequent immunoblot analysis. Immunoblotting Cell pellets were lysed in a modified EBC lysis buffer (50mM Tris-Cl pH 8.0, 250 mM NaCl, 0.5% NP-40, 5 mM EDTA) supplemented with a protease inhibitor cocktail (Complete, Roche Applied Science, #11836153001) and phosphatase inhibitors (PhosSTOP Sigma #04906837001). Soluble cell extracts were quantified using the Bradford Protein Assay. 20 µg of protein per sample was boiled after adding 3X sample buffer (6.7% SDS, 33% Glycerol, 300 mM DTT, and Bromophenol Blue) to a final concentration of 1X, resolved by SDS-PAGE using either 10% or 8% SDS-PAGE, semi-dry transferred onto nitrocellulose membranes, blocked in 5% milk in Tris-Buffered Saline with 0.1%Tween 20 (TBS-T) for 1h, and probed with the indicated primary antibodies overnight at 4°C. Membranes were then washed three times in TBS-T, probed with the indicated horseradish peroxidase conjugated (HRP) secondary antibodies for 1h at room temperature, and washed three times in TBS-T. Bound antibodies were detected with enhanced chemiluminescence (ECL) western blotting detection reagents (Immobilon, Thermo Fisher Scientific, #WBKLS0500) or Supersignal West Pico (Thermo Fisher Scientific, #PI34078). The primary antibodies and dilutions used were: Rabbit anti-EED (Cell Signaling #85322S, 1:1000), rabbit anti-ASCL1 (Abcam #Ab211327, 1:1000), rabbit anti-NEUROD1 [EPR4008] (Abcam #Ab109224, 1:1000), rabbit anti-RAS (E8N8L) (Cell Signaling #67648, 1:1000), rabbit anti-Rb1 (Abcam #181616, 1:2000), rabbit rodent specific anti-p53 (D2H9O) (Cell Signaling Technology #32532, 1:1000), rabbit anti-p130 (Abcam #Ab76234, 1:1000), rabbit anti-Cre Recombinase (D7L7L) (Cell Signaling Technology #15036, 1:1000), Mouse Anti-Cas9 (Cell Signaling Technology #14697, 1:1000), mouse anti-β-actin (clone AC-15) (Sigma, #A3854, 1:25,000), mouse α-Vinculin (hVIN-1) (Sigma, # V9131, 1:10000). Histone extractions were performed as described previously 26 with the following primary antibodies: Rabbit anti-Histone H3 (D1H2) (Cell Signaling Technology #4499S, 1:1000) and rabbit anti-Tri-Methyl-Histone H3 (Lys27) (Cell Signaling Technology #9733S, 1:1000). The secondary antibodies and dilutions used were: Goat Anti-Mouse (Jackson ImmunoResearch #115-035-003) and Goat anti-Rabbit (Jackson ImmunoResearch #111-035-003) and used at 1:5000. Bulk RNA-Sequencing and Analysis Tumors were harvested at necropsy and were flash-frozen. RNA was extracted using Quick-RNA™ Miniprep kit (Cat # R1055, Zymo Research, CA, USA) including a DNase digestion step according to the manufacturer’s instructions and RNA sequencing was performed as described below. Total mRNA samples in each experiment were submitted to Novogene Inc. The libraries for RNA-seq are prepared using NEBNext Ultra II non-stranded kit. Paired end 150bp sequencing was performed on Novaseq6000 sequencer using S4 flow cell. Sequencing reads were mapped to the mm10 genome by STAR. Statistics for differentially expressed genes were calculated by DESeq2 (1.36.0). Bulk RNA-sequencing heatmaps were generated by calculating z score from log transformed FPKM values. To calculate z score, a data point was subtracted the mean and then divide the result by the standard deviation. Transcriptomic Profile Correlation Analysis Transcriptomic profiles of never-transformed and pre-transform LUAD and transformed and de novo SCLC [Transcript Per Million (TPM) values from RNA-seq data] were obtained from published work by Quintanal-Villalonga et al. 2021 5 and top 100 signature genes from each group vs. all others were determined using log-transformed TPM values with the eBays function in the limma package (3.56.2) in R (4.3.1). Log transformed TPM of signature genes were merged with EED -isogenic RPR2 bulk tumor RNA-seq data, Pearson correlations were calculated with the cor function in base stats package in R (4.3.1). Correlation coefficients were further normalized by calculating z score as described above. Transcriptomic Data Analysis of the Cancer Genome Atlas (TCGA) Batch normalized mRNA expression, mutation, and copy number alteration data from Lung adenocarcinoma (LUAD) patient tumors in the TCGA dataset were downloaded from cBioPortal ( cbioportal.org ) 68 . Samples with either mutation or deep deletion (copy number alteration value=-2) in RB1 and/or TP53 were categorized as a group and data was analyzed for all LUADs or only LUADs harboring EGFR mutations. Statistical testing between groups were performed using Tukey test to adjust for multiple comparisons. Gene Set Enrichment Analysis GSEA software was downloaded from the Gene Set Enrichment Analysis website [ http://www.broad.mit.edu/gsea/downloads.jsp ] 70 . Pre-ranked GSEA was performed using hallmark gene sets from the human MSigDB Collections 32 in Figs. 2D , 3C ,4M,S9L, top 100 neuroendocrine genes, ASCL1 target genes 59 in Fig. 2D , human and mouse mucinous lung tumor signature 36 in Fig. 2I , and ASCL1 correlated genes 59 , NEUROD1 correlated genes 59 , IMPOWER133-SCLC-N (NMF1) and SCLC-A (NMF2) gene signatures 12 in Fig. 3E . The Gene Set Variation Analysis (GSVA) method from the GSVA Bioconductor package (version 1.50.5), was used to calculate enrichment scores from tuft cell marker gene lists (Fig. S4B) and top 100 NE gene list (Fig. S4C) using normalized read counts from bulk EED -isogenic RPR2 tumor RNA-seq data and from gene signatures of MAPK targets 67 , neuroendocrine markers 12 , tumor proliferation rate 12 , hallmark_MYC, hallmark_G2M, and neutrophil markers 12 and human and mouse mucinous lung tumor signature ( Fig. 6H-I ) using normalized read counts from bulk RNA-seq data from EGFR -Mutant CRP and ERP GEMM lung tumors. GSVA was also performed using functional EED target genes on normalized read counts from IMPOWER133 NE and non-NE SCLC bulk RNA-seq data as previously reported 12 ( Fig. 3R , 5E ). For gene set over-representation analysis, candidate gene sets were tested if over-represented against the human hallmark pathways from the MSigDB in Figs. 3Q , 5D , S8E, S9K,L, and S10I,L, mouse KEGG pathway collection 71 in Figs. 5C,G ,N, 7E,K, S7D,G, S9I, and S10K using R package Clusterprofiler (4.8.3). Mouse gene atlas 72 ( Fig. 2H ), Gene Ontology Biological Processes 73 (Fig. S10C,D), and ENCODE and ChEA consensus TFs 74 ( Figs. 2B , 7I , S7C,F, S10A,B) using R package enrichR (3.4). Reverse-Transcriptase Quantitative PCR (RT-qPCR) RNA was extracted using Quick-RNA™ Miniprep kit (Cat#R1055, Zymo Research, CA, USA) according to the manufacturer’s instructions. RNA concentration was determined using the Nanodrop 8000 (Thermofisher Scientific). A cDNA library was synthesized using iScript Reverse Transcription Supermix for RT-qPCR (Biorad #1708841) according to the manufacturer’s instructions. qPCR were performed using the SsoAdvanced Universal SYBR Green Supermix (Biorad #1725271) according to the manufacturer’s instructions. The ΔΔC T Method was used to analyze data. The C T values for each primer set were then normalized to the C T value of 36b4 . The data from Fig. 4C was then normalized to the control to determine the relative fold change in mRNA expression. The following probes were used for qPCR with SYBR Green: Mouse Neurod1 Forward (5’-AGGCTCCAGGGTTATGAGATCG-3’), Mouse Neurod1 Reverse (5’-TGAGAACTGAGACACTCATCTG-3’), Mouse Met Forward (5’-GACCTTAAGCGAGAGCACGA-3’), Mouse Met Reverse (5’-ATGCACTGTATTGCGTCGTC-3’), Mouse 36b4 Forward (5’- CTGTTGGCCAATAAGGTGCC-3’), Mouse 36b4 Reverse (5’- GTTCTGAGCTGGCACAGTGA-3’). Immunohistochemistry Immunohistochemistry was performed on the Leica Bond III automated staining platform using the Leica Biosystems Refine Detection Kit (Leica; DS9800). FFPE tissue sections were baked for 30 minutes at 60°C and deparaffinized (Leica AR9222) prior to staining. Primary antibodies were incubated for 30 minutes, visualized via DAB, and counterstained with hematoxylin (Leica DS9800). The slides were rehydrated in graded alcohol and coverslipped using the HistoreCore Spectra CV mounting medium (Leica 3801733). 4 µm-thick sections were cut from formalin-fixed paraffin-embedded mouse tumor samples to perform single IHC studies using antibodies recognizing the following antigens: ASCL1 (rabbit monoclonal antibody clone EPR19840, Abcam; 1:100 concentration, antigen retrieval with citrate for 30 min.); NEUROD1 (rabbit monoclonal antibody clone EPR20766, Abcam; 1:100 concentration, antigen retrieval with citrate for 30 min.); POU2F3 (rabbit polyclonal antibody, Atlas Antibodies; 1:400 concentration, antigen retrieval with citrate for 30 min.); EED (rabbit monoclonal antibody clone E4L6E, Cell Signaling Technology; 1:200 concentration, antigen retrieval with EDTA for 30 min.); EGFR mutant-specific L858R antibody (rabbit monoclonal antibody clone 43B2, Cell Signaling Technology; 1:100 concentration, antigen retrieval with citrate for 30 min.); tri-methyl-histone H3 (Lys27) (rabbit monoclonal antibody clone C36B11, Cell Signaling Technology; 1:100 concentration, antigen retrieval with citrate for 30 min.); and Synaptophysin (rabbit monoclonal antibody clone SP11, Invitrogen; 1:50 concentration, antigen retrieval with citrate for 30 min.). The immunostained slides were scanned at 40x magnification using the PhenoImager HT slide imager (Akoya Biosciences). For all markers, visual assessment of positive and negative cells was done by two independent pathologists (YNL, GRO). To quantify NEUROD1-positive tumor cells, tumor areas on each slide were identified and manually annotated by two pathologists (YNL, GRO) using the HALO image analysis platform (Indica Labs). The HALO multiplex-IHC v3.2.5 algorithm (Indica Labs) was then used to quantify the number of positive and negative tumor cells, and the percentage of NEUROD1-positive tumor cells was subsequently calculated. ATAC-sequencing Nuclei isolation for ATAC-sequencing from mouse tumor tissue was performed using ATAC-Seq Kit (ActiveMotif, cat# 53150) following manufacture’s instruction. Briefly, 30mg frozen mouse RPR2 lung tumor tissues were minced with a razor blade and homogenized in ATAC-seq Lysis Buffer in a dounce homogenizer. Lysates were then filtered through a 40 micron cell strainer and Trypan blue at 0.4% was used to count 100,000 nuclei which were subsequently rinsed with ice-cold PBS and resuspended in ice-cold ATAC-seq Lysis Buffer. After mixing with Tagmentation Master Mix, tagmentation reactions were incubated at 37 °C and shaken at 800 RPM for 30 minutes on a thermomixer. Transposed DNA was purified using columns and libraries were amplified following manufacture’s manual. 50 million paired-end reads, where each read is 150 base pairs long, were sequenced on a NextSeq instrument (Illumina). ATAC-seq Data Analysis All samples were processed through the computational pipeline developed at the Dana-Farber Cancer Institute Center for Functional Cancer Epigenetics (CFCE) using primarily open-source program CHIPS 75 , 76 . Sequence tags were aligned with Burrows-Wheeler Aligner (BWA) 77 to build mm10 and uniquely mapped, non-redundant reads were retained. These reads were used to generate binding sites with Model-Based Analysis of ChIP-Seq 2 (MACS v2.1.1.20160309), with narrow peak option and a q-value (FDR) threshold of 0.01 78 . Average peak tracks were calculated with bigwigAverage function in deeptools (3.5.1) and visualized by IGV (v2.14.1) 79 . Peaks from all samples were processed using the DiffBind pipeline (v3.18) 80 for differential binding analysis. Briefly, peaks were merged to create a union set of sites. Sequencing depth normalization was performed on each sample and DEseq2 was used to determine differential peaks. Log-transformed fold changes were subsequently shrunk with lfcshrunk for more accurate foldchange estimation. Differential binding peaks from each group were used for motif analysis by the motif search findMotifsGenome.pl in HOMER (v3.0.0) 81 , with cutoff q-value≤ 1e-10. The signals of each sample on differential binding sites were visualized by deeptools 82 . For principal component analysis, all peaks were used in the analysis and graphic visualization. Chromatin Immuno-Precipitation Sequencing (ChIP-seq) Analysis of De Novo SCLC and EGFR -Mutant LUAD Patient Derived Xenograft (PDX) Tumors ChIP-seq for H3K27me3 and H3K4me3 from 4 NE de novo SCLC and 5 EGFR -Mutant LUAD PDX samples were previously published 52 and the fastq files were obtained from the Gene Expression Ominibus (GEO accession#: GSE269746). High quality reads passing quality control using fastp were aligned to HumanG1Kv37 reference with BWA 83 . Uniquely mapped, non-redundant reads were generated by picard tools (2.19.0) and samtools (1.9) and subsequently used to generate binding sites with MACS2 (MACS v2.1.1.20160309) with q-value (FDR) threshold of 0.01 and narrow peak option for H3K4me3 and broad peak option for H3K27me3 78 . Peaks from all samples were processed using the DiffBind pipeline (v3.18) 80 for differential binding analysis as described above. Mouse EED -Isogenic RPR2 GEMM Tumors ChIP-seq Sample Preparation and Sequencing Frozen mouse tumor tissue was dry pulverized in liquid nitrogen until it was finely powdered. Pulverized tissue was resuspended for crosslinking in 2 mM of disuccinimidyl glutarate (DSG, Pierce) for 45 minutes at room temperature. The crosslinked material was pelleted by centrifugation at 2500 RCF for 5 min and resuspended in 1 ml of 1% formaldehyde for 10 minutes. Formaldehyde was quenched with 0.125 M glycine for 5 minutes at room temperature and then the crosslinked material washed with PBS. Washed cells were now pelleted and resuspended in 500 μl of 1% SDS (50 mM Tris-HCl pH 8, 10 mM EDTA) and sonicated for 10 minutes using Covaris E220 sonicator at the following settings: 140 peak incident power, 5% duty factor and 200 cycles per burst) in 1 ml AFA fiber millitubes. Chromatin was immunoprecipitated with 5 μg of antibody (EED Cell Signaling E4L6E; H3K27ac Diagenode C15410196; H3K27me3 Active Motif 61018; H3K4me3 Abcam ab8580). 5 μg of chromatin was used for histone mark ChIPs, and 40 μg of chromatin was used for EED ChIPs. Imunoprecipitated Chromatin was pulled down using Protein A/G beads (Invitrogen Dynabeads 1000-2D and 1000-4D). Chromatin antibody complex immobilized on beads was washed 6X with RIPA buffer and eluted in 1% SDS and 0.1M Sodium bicarbonate. ChIP-ed DNA was crosslinked with Proteinase K and column purified. ChIP-seq libraries were made using Swift DNA Library Prep Kit (Swift Biosciences 10024). 75bp paired end reads were sequenced on a NextSeq instrument (Illumina). ChIP-seq Analysis and Peak Calling of Mouse EED -Isogenic RPR2 GEMM Tumors Peak calling and data analysis All samples were processed through the computational pipeline developed at the Dana-Farber Cancer Institute Center for Functional Cancer Epigenetics (CFCE) using primarily open-source programs 75 , 76 . Sequence tags were aligned with Burrows-Wheeler Aligner (BWA) 83 to build mm9 and uniquely mapped, non-redundant reads were retained. These reads were used to generate binding sites with Model-Based Analysis of ChIP-Seq 2 (MACS v2.1.1.20160309), with a q-value (FDR) threshold of 0.01 78 . We evaluated multiple quality control criteria based on alignment information and peak quality: (i) sequence quality score; (ii) uniquely mappable reads (reads that can only map to one location in the genome); (iii) uniquely mappable locations (locations that can only be mapped by at least one read); (iv) peak overlap with Velcro regions, a comprehensive set of locations – also called consensus signal artifact regions – in the genome that have anomalous, unstructured high signal or read counts in next-generation sequencing experiments independent of cell line and of type of experiment; (v) number of total peaks (the minimum required was 1,000); (vi) high-confidence peaks (the number of peaks that are tenfold enriched over background); (vii) percentage overlap with known DHS sites derived from the ENCODE Project (the minimum required to meet the threshold was 80%); and (viii) peak conservation (a measure of sequence similarity across species based on the hypothesis that conserved sequences are more likely to be functional). Differential binding analyses Peaks from all samples were merged to create a union set of sites for each transcription factor and histone mark using bedops 84 . Sample-sample correlation and differential peaks analysis were performed by the CoBRA pipeline 76 . Read densities were calculated for each peak for each sample and used for the comparison of cistromes across samples. Sample similarity was determined by hierarchical clustering using the Spearman correlation between samples. differential peaks were identified by DEseq2 with adjusted P ≤ 0.05. A total number of reads in each sample was applied to the size factor in DEseq2, which can normalize the sequencing depth between samples. Peaks from each group were used for motif analysis by the motif search findMotifsGenome.pl in HOMER (v3.0.0) 81 , with cutoff q-value ≤ 1e-10. Single Cell RNA-Sequencing Sample Preparation The single-cell RNA sequencing (scRNA-seq) experiments in autochthonous CRISPR-based EED -WT or EED -Mutant RPR2 GEMM models were performed as previously described 85 . Briefly, 3-4 month old male and female homozygous BL6J LSL-Cas9 mice (Jackson No. 026175) were intratracheally injected with EED -Mutant or EED -WT adenoviruses (see adenovirus method above). Once mice became symptomatic from their tumors (see method above), 6 independent EED- Mutant mice (4 EED -Mutant RPR2 LUADs and 2 EED -Mutant RPR2 SCLCs) and 3 independent EED WT mice were euthanized and lung tumors dissected and finely minced mechanically using a razor blade and then enzymatically digested with Mouse Tumor Dissociation Kit (Miltenyi Biotec, #130-096-730) following the manufacturer’s instruction. Briefly, minced tumor tissue was transferred to a gentleMACS C Tube containing enzyme mix prepared with 20% of Enzyme R option to preserve cell surface epitopes. Dissociation using the gentleMACS Octo Dissociator with Heaters (Miltenyi Biotec, #130-096-427) was performed using the 37C_m_TDK_2 gentleMACS Program. The single cell suspensions were resuspended in RPMI containing 10% FBS and subsequently passed through a 70 μm Cell Strainer (Greiner, #542070) and centrifuged at 300 x g for 3 min followed by 2 washes with 0.04% UltraPure Bovine Serum Albumin (Invitrogen, AM2616) in DPBS. Finally, dissociated cells were resuspended in DPBS with 0.04% UltraPure BSA and cell counts were measured with a Vi-CELL XR Cell Viability Analyzer (Beckman Coulter). Cells were then diluted in 0.04% BSA/DPBS at a cell concentration of 1000 cells/µL. About 16,000 cells were loaded onto a 10x Genomics Chromium TM instrument (10x Genomics) according to the manufacturer’s instructions. The scRNA-seq libraries were processed using Chromium Next GEM Single Cell 5’ Kit v2 kit (10x Genomics). Quality controls for amplified cDNA libraries and final sequencing libraries were performed using Bioanalyzer High Sensitivity DNA Kit (Agilent). The sequencing libraries for scRNA-seq were normalized to 4 nM concentration and pooled. The pooled sequencing libraries were sequenced on Illumina NovaSeq S4 300 cycle platform. The sequencing parameters used were: Read 1 of 26bp, Read 2 of 90bp, Index 1 of 10bp and Index 2 of 10bp. EED WT RPR2 scRNA-seq data from 3 independent mice were previously published 85 . Single-Cell RNA Sequencing Data Analysis Custom reference genome was established by adding EGFP and Cas9 sequence 86 to the mouse genome (mm10-2020-A). Cell ranger version 6.0.2 pipeline (10x Genomics) was used to align sequencing data to the custom mouse reference genome and generate the gene-level counts matrix. Unfiltered raw counts data was imported into Seurat v4 R package (version 4.1.0) 87 for downstream processing and analysis. Low quality cells were filtered out using following thresholds: total UMI counts < 500, number of transcripts < 350, log 10 TranscriptsPerUMI <= 0.8, and cells with more than 15% transcripts mapping to mitochondrial genes. In addition, genes expressed in less than ten cells were removed. The UMI counts matrices were then natural-log normalized and scaled with Seurat’s ‘NormalizeData’ and ‘ScaleData’ functions. Single Nuclei RNA-Sequencing Sample Preparation The single-nuclei RNA sequencing (snRNA-seq) sample preparation from EED -Mutant RPR2 LUAD tumors were performed using Chromium Nuclei Isolation Kit with RNase Inhibitor (10x Genomics, PN-1000494) following manufacture’s instruction (10x Genomics, CG000124). Briefly, 30 mg snap-frozen tumor tissue from 3 independent EED -Mutant RPR2 mice was dissociated in Lysis Buffer with pestle and centrifuged for 20s, in 16,000 rcf under 4 °C to a column to collect lysed nuclei in the flowthrough. Debris in the flowthrough containing nuclei was removed by Debris Removal Buffer before nuclei were washed 2 times with 5 min, 500 rcf, 4 °C spin. Nuclei was then counted using ViaStain AOPI Staining Solution (Revvity, CS2-0106-5ML) on a Nexcelom Cellometer K2 Fluorescent Cell Counter (Revvity, CMT-K2-MX-150) and resuspended in 1,000 nuclei/ul DPBS with 0.04% ultra-pure Bovine Serum Albumin (Invitrogen, AM2616). About 20,000 nuclei were loaded onto a ChromiumTM X instrument (10x Genomics) according to the manufacturer’s instructions. The snRNAseq libraries were processed using Chromium GEM-X Single Cell 5’ Kit v3 (10x Genomics). Quality controls for amplified cDNA libraries and final sequencing libraries were performed using Bioanalyzer High Sensitivity DNA Kit (Agilent). Libraries were subsequently quantified by Qubit fluorometer and Agilent 4200 TapeStation system. Library pooling and indexing was evaluated with shallow sequencing on an Illumina MiSeq. Subsequently, libraries were sequenced by the Molecular Biology Core Facilities at Dana-Farber Cancer Institute on an Illumina NovaSeq X Plus with the following sequencing parameters: 28bp, 10bp, 10bp, 90bp. Single-Nuclei RNA Sequencing Data Analysis Custom reference genome was established by adding EGFP and Cas9 sequence 86 to the mouse genome (mm10-2020-A). Cell ranger version 9.0.0 pipeline (10x Genomics) was used to align sequencing data to the custom mouse reference genome and generate the gene-level counts matrix. Unfiltered raw counts data was imported into Seurat v4 R package (version 4.1.0) 87 for downstream processing and analysis. Low quality cells were filtered out using following thresholds: total UMI counts < 500, number of transcripts < 375, log 10 TranscriptsPerUMI = 0.2 for scRNA-seq and >= 0.1 for snRNA-seq. In addition, genes expressed in less than five cells were removed. The UMI counts matrices were then natural-log normalized and scaled with Seurat’s ‘NormalizeData’ and ‘ScaleData’ functions. Dimension Reduction, Cluster Analysis and Visualization of scRNA-Seq and snRNA-Seq Data The Seurat v4 R package was used for dimension reduction and clustering. Top 2,000 genes with the highest variance were selected using the ‘FindVariableFeatures’ function with ‘vst’ method to perform linear dimensional reduction (principal component analysis) using the ‘RunPCA’ function, and top 40 principal components for scRNA-seq and top 25 principal components for snRNA-seq were used to perform graph-based unsupervised clustering with the ‘FindNeighbors’ and ‘FindClusters’ functions and Uniform Manifold Approximation and Projection for Dimension Reduction (UMAP) ( arXiv:1802.03426 ) for data visualization in two-dimensional space. Automatic cluster annotation using the R package SingleR (version 1.8.1) 88 with ImmGenData 89 , 90 and manual tumor cluster annotation with markers of EGFP and Cas9 were performed to select the tumor population. Subsequently, Ptprc positive tumor cells, Cas9 and EGFP positive tumor microenvironment were filtered out. Then, the tumor subpopulations were reanalyzed. Seurat’s ‘CellCycleScoring’ function and S phase and G2/M phase specific gene lists were used to calculate the G2M and S phase gene expression scores and to be regressed out using the ‘var.to.regress’ option in the ‘SCTransform’ function with default parameters. Top 3,000 genes excluding mitochondrial and ribosomal genes with the highest variance were selected using the ‘FindVariableFeatures’ function with ‘vst’ method to perform linear dimensional reduction (principal component analysis) using the ‘RunPCA’ function, and top 40 principal components were used to plot UMAP for data visualization in two-dimensional space in scRNA-seq and top 9 principal components were used in snRNA-seq. Differential expression profile were obtained with Seurat’s ‘FindMarkers’ function using test.use = “wilcox” to compare differentially expressed genes. Cell trajectory analysis was performed on tumor populations with the Monocle2 package (V2.24.0). For snRNA-seq, top 100 of each of differentially upregulated and downregulated genes between the Neurod1 + cluster and Muc13 + cluster were used as the ordering genes. For scRNA-seq, top 300 differentially expressed genes between EED -Mutant RPR2 LUAD tumor cells and EED -WT RPR2 SCLC tumor cells were used as ordering genes. To predict cell biological trajectories, Monocle2 estimated size factor and dispersion, followed by reduceDimension function with the “DDRTree” method to reduce the high-dimensional space as a way to infer cells progression through a given biological process. For snRNA-seq, top 300 non-mitochondrial genes that significantly correlates to the trajectory (based on descending order of q value from “differentialGeneTest” function) were plotted in the heatmap in Fig. 4H . Tumor and its Microenvironment Communication Analysis using Differential NicheNet To study communication between the tumor and cells of the microenvironment in EED -WT SCLC, EED -Mutant SCLC, and EED -Mutant LUAD tumors, we used the nichenetr package V 1.1.0.39. NicheNet predicts which differentially expressed ligands from one tumor microenvironment population are most likely to affect target gene expression of tumor population and which specific target genes are affected by which of these predicted ligands. It uses curated ligand-receptor and ligand-target databases 42 as references to evaluate the cellular communication networks between two cell types. Human gene names were converted to their one-to-one orthologs using ‘convert_human_to_mouse_symbols’ function. Ligands, receptors, and targets that expressed in at least 10% (default cutoff) of the single cells in a specific cluster were included. Differential genesets of interest were calculated using a log transformed fold change threshold of 0.15 (recommended cutoff for 10x datasets), and 2889 genes were included in EED -WT SCLC and 1668 genes of interest were included in EED -Mutant LUAD. Next, ligand-receptor-target links were prioritized using the weights of scaled properties of interest as follows: “scaled_ligand_score” = 5, “scaled_ligand_expression_scaled” = 1, “ligand_fraction” = 1, “scaled_ligand_score_spatial” = 0, “scaled_receptor_score” = 0.5, “scaled_receptor_expression_scaled” = 0.5, “receptor_fraction” = 1, “ligand_scaled_receptor_expression_fraction” = 1, “scaled_receptor_score_spatial” = 0, “scaled_activity” = 0, “scaled_activity_normalized” = 1, “bona_fide” = 1. Confident Tumor-TME interactions that are significantly enriched in EED -Mutant LUAD were further identified by significantly different expression of ligand in TME and receptor in tumors with Bonafide interaction. Tumor-expressing receptors in significantly enriched interaction pairs were further overlapped with significantly upregulated genes (p value<0.05) in mouse SCLC RPR2 cell line 1014 after EED genetic inactivation. Scaled receptor score (Log2FoldChange comparing EED -Mutant LUAD and EED -WT SCLC) of unique receptors that are 1. Directly activated by loss of EED, 2, predicted activated by ligand from TME only in EED -Mutant LUAD, were plotted in Fig. 4Q . CNV Inference and Phylogenetic Tree Construction R package inferCNV (1.16.0) was used to detect copy number variations from snRNA-seq data. Normal pulmonary epithelium and stroma cells were used as reference for the tumor population from the corresponding sample. The following parameters in the “run” function of infercnv were used: cutoff = 0.1, denoise = TRUE, cluster_by_groups = T, HMM = T. All genes from HMM copy number inferences (from file 17_HMM_predHMMi6.hmm_mode-samples.pred_cnv_genes.dat) were used to construct the phylogenetic tree; copy number was set to diploid if a cluster did not have an inferred CNV event predicted. Phylogenetic trees were inferred using the Maximum Parsimony (MP) method with the Subtree-Pruning-Regrafting (SPR) algorithm 91 with search level 1 in MEGA11 92 . Flow Cytometry Fresh single cell suspensions from EED WT or EED -Mutant RPR2 mice tumors were prepared, washed twice in PBS, resuspended in FACS buffer (D-PBS containing 2% FBS) and transferred in 1.5 mL Eppendorf tubes. After washing, cells were incubated with fluorophore conjugated anti-H-2Ld/H-2Db (Biolegend #114507, clone 28-14-8) or the isotype control (Biolegend #400212, IgG2a, κ isotype Ctrl) at 1:100 dilution in the dark for 30 minutes at room temperature. After washing, cells were resuspended in FACS buffer, transferred to flow cytometry tubes containing a 40 μm filter and analyzed on a LSR Fortessa flow cytometer (Becton Dickinson, Franklin Lakes, NJ). Data analyses were performed with FlowJo software (V10). DNA Sequencing of GEMM Tumors Genomic DNA from the tumors indicated was isolated using the QIAamp DNA Blood Mini Kit (Qiagen #51106) according to the manufacturer’s instructions. PCR was done by using KOD Xtreme™ Hot Start DNA Polymerase (Sigma-Aldrich, Cat# 71975) and following set of primers to amplify the genomic region of Eed targeted by sg Eed #2: forward Eed #2 (5’- TACTGCTCACAGGACGAT -3’), reverse Eed #2 (5’- GCTTCCAGCACATACCTT -3’). Two consecutive PCRs were used for Rb1 where a 731 bp product was generated from the first PCR and subsequently used as template for the second PCR with the following sets of primers: 1 st : forward Rb1 (5’- CTATCATCTTCATGCTACAA -3’), reverse Rb1 (5’- GCAGAATAAAATTCTACCAGG -3’), 2 nd : forward Rb1 (5’- GCATATATATCTACTTCAGCTG -3’), reverse Rb1 (5’- GGTCAATGTGGAATACACAATTG -3’).The final PCR product was then column purified using Qiagen’s gel extraction kit and sequenced using next-generation Amplicon sequencing by the MGH DNA Core Facility or Sanger Sequencing (Azenta Life Sciences). EED cDNA for Exogenous EED Re-Expression pTwist Lenti SFFV Puro WPRE lentiviral vectors encoding: 1) EED WT-HA or 2) EED Triple Mutant-HA (F97A/Y148A/Y365A) which encodes a functionally inactive EED Mutant-HA were codon optimized and synthesized by Twist Biosciences. All plasmids were sequence verified by sanger sequencing. The functionally inactive EED Mutant was confirmed as was unable to restore H3K27me3 upon re-expression relative to EED-WT which restored H3K27me3 upon re-expression. Statistics and Reproducibility For all GSEA, enrichment analysis from RNA-sequencing, ATAC-sequencing and ChIP-sequencing data, statistical significance was calculated corrected for multiple hypothesis testing. For survival analysis in Fig. 1B , 6C , log-rank test was used. For correlation analysis in Fig. 2E , pearson correlation was calculated. For the GREAT analysis of ATAC-sequencing data in Figs. S8C,D, S10F,H, a binomial p-value was calculated as described previously 93 . For all other experiments, statistical significance was calculated using unpaired, two-sided Students t-test. p -values were considered statistically significant if the p -value was <0.05. Error bars represent S.E.M. unless otherwise indicated. For all experiments with statistical data, the number of independent biological experiments are described in each figure legend. For immunoblot analyses in Figs. 3K,M , S6A,D-F, S7A, at least 3 biological independent experiments were performed and representative immunoblots are shown. For immunoblot analyses in Figs. 4K , S9C, immunoblots contains multiple independent tumors. For immunohistochemistry experiments and H&E staining, representative micrographs were shown from independent tumors from independent mice. For the mouse experiments in Fig. 1 , 48 mice were included and mice were completely randomized to receive sgEED RPR2 or sgControl RPR2 adenoviruses. For the mouse experiments in Fig. 6 , 59 mice were included and mice were completely randomized to receive sgEED RP or sgControl RP adenoviruses. For all sequencing experiments (RNA-seq, ATAC-seq, scRNA-seq, snRNA-seq and ChIP-seq) the investigators that processed and sequenced the samples [DFCI’s CFCE core (ChIP-seq), DFCI MBCF core (snRNA-seq and ATAC-seq), DFCI TIGL core (scRNA-seq), or Novogene (RNA-seq)] were blinded to the identity of the samples. For all cell culture experiments, each experiment was repeated in at least 3 biological independent experiments as specified in the figure legend. No statistical methods were used to pre-determine sample sizes, but our sample sizes are similar to those reported in previous publications 25 , 26 . Data distribution was assumed to be a normal distribution, but this was not formally tested. For cell culture experiments, blinding was not possible. For all experiments, no data were excluded from any analyses. Acknowledgements This work was supported by a William Raveis Charitable Fund Damon Runyon Clinical Investigator Award (CI-101-19, M.G.O.), a DF/HCC SPORE CEP Award (P50CA265826, M.G.O.), an NCI P01 award (P01CA295524, M.G.O.), an American Cancer Society Postdoctoral Fellowship (PF-24-1252572-01-CCB, Y.L.), the Kaplan Family Fund (M.G.O), and an NCI R01 award (R01CA263715, K.P.). S.L. is supported by the NCI Research Specialist Award (R50CA251956). We thank Dr. David Barbie for thoughtful discussions and members of the Oser, Barbie, and Janne labs for their critical feedback. We thank the DFCI Molecular Biology Core Facility including Zach Herbert and Maura Berkeley who used an Illumina NovaSeq X Plus that was purchased with funding from a National Institutes of Health SIG grant 1S10OD036228-01 for this work. We also thank the Dana-Farber/Harvard Cancer Center for the use of the Specialized Histopathology Core, which provided histology and immunohistochemistry service. Dana-Farber/Harvard Cancer Center is supported in part by an NCI Cancer Center Support Grant # NIH 5 P30 CA06516. The results in Figs. 7O-Q , S10M-O are based upon data generated by the TCGA Research Network: https://www.cancer.gov/tcga . Funder Information Declared William Raveis Charitable Fund Damon Runyon Clinical Investigator Award , CI-101-19 National Cancer Institute , P50CA265826 , P01CA295524 , R01CA263715 , R50CA251956 American Cancer Society, https://ror.org/02e463172 , PF-24-1252572-01-CCB Kaplan Family Fund References ↵ Rudin , C. M. , Brambilla , E. , Faivre-Finn , C. & Sage , J. Small-cell lung cancer . Nat Rev Dis Primers 7 , 3 , doi: 10.1038/s41572-020-00235-0 ( 2021 ). OpenUrl CrossRef PubMed ↵ Oser , M. G. , Niederst , M. J. , Sequist , L. V. & Engelman , J. A . Transformation from non-small-cell lung cancer to small-cell lung cancer: molecular drivers and cells of origin . Lancet Oncol 16 , e165 – 172 , doi: 10.1016/S1470-2045(14)71180-5 ( 2015 ). OpenUrl CrossRef PubMed ↵ Quintanal-Villalonga , A. et al. Lineage plasticity in cancer: a shared pathway of therapeutic resistance . Nat Rev Clin Oncol 17 , 360 – 371 , doi: 10.1038/s41571-020-0340-z ( 2020 ). OpenUrl CrossRef PubMed ↵ Niederst , M. J. et al. RB loss in resistant EGFR mutant lung adenocarcinomas that transform to small-cell lung cancer . Nat Commun 6 , 6377 , doi: 10.1038/ncomms7377 ( 2015 ). OpenUrl CrossRef PubMed ↵ Quintanal-Villalonga , A. et al. Multiomic Analysis of Lung Tumors Defines Pathways Activated in Neuroendocrine Transformation . Cancer Discov 11 , 3028 – 3047 , doi: 10.1158/2159-8290.CD-20-1863 ( 2021 ). OpenUrl Abstract / FREE Full Text Sequist , L. V. et al. Genotypic and histological evolution of lung cancers acquiring resistance to EGFR inhibitors . Sci Transl Med 3 , 75ra26 , doi: 10.1126/scitranslmed.3002003 ( 2011 ). OpenUrl Abstract / FREE Full Text ↵ Marcoux , N. et al. EGFR-Mutant Adenocarcinomas That Transform to Small-Cell Lung Cancer and Other Neuroendocrine Carcinomas: Clinical Outcomes . J Clin Oncol 37 , 278 – 285 , doi: 10.1200/JCO.18.01585 ( 2019 ). OpenUrl CrossRef PubMed ↵ Offin , M. et al. Concurrent RB1 and TP53 Alterations Define a Subset of EGFR-Mutant Lung Cancers at risk for Histologic Transformation and Inferior Clinical Outcomes . J Thorac Oncol 14 , 1784 – 1793 , doi: 10.1016/j.jtho.2019.06.002 ( 2019 ). OpenUrl CrossRef PubMed ↵ Davies , A. , Zoubeidi , A. , Beltran , H. & Selth , L. A . The Transcriptional and Epigenetic Landscape of Cancer Cell Lineage Plasticity . Cancer Discov 13 , 1771 – 1788 , doi: 10.1158/2159-8290.CD-23-0225 ( 2023 ). OpenUrl CrossRef PubMed ↵ Gardner , E. E. et al. Lineage-specific intolerance to oncogenic drivers restricts histological transformation . Science 383 , eadj1415 , doi: 10.1126/science.adj1415 ( 2024 ). OpenUrl CrossRef PubMed ↵ George , J. et al. Comprehensive genomic profiles of small cell lung cancer . Nature 524 , 47 – 53 , doi: 10.1038/nature14664 ( 2015 ). OpenUrl CrossRef PubMed ↵ Nabet , B. Y. et al. Immune heterogeneity in small-cell lung cancer and vulnerability to immune checkpoint blockade . Cancer Cell 42 , 429 – 443 e424 , doi: 10.1016/j.ccell.2024.01.010 ( 2024 ). OpenUrl CrossRef PubMed ↵ Sivakumar , S. et al. Integrative Analysis of a Large Real-World Cohort of Small Cell Lung Cancer Identifies Distinct Genetic Subtypes and Insights into Histologic Transformation . Cancer Discov 13 , 1572 – 1591 , doi: 10.1158/2159-8290.CD-22-0620 ( 2023 ). OpenUrl CrossRef PubMed ↵ Lee , J. K. et al. Clonal History and Genetic Predictors of Transformation Into Small-Cell Carcinomas From Lung Adenocarcinomas . J Clin Oncol 35 , 3065 – 3074 , doi: 10.1200/JCO.2016.71.9096 ( 2017 ). OpenUrl CrossRef ↵ Zhang , B. et al. Brief Report: Comprehensive Clinicogenomic Profiling of Small Cell Transformation From EGFR-Mutant NSCLC Informs Potential Therapeutic Targets . JTO Clin Res Rep 5 , 100623 , doi: 10.1016/j.jtocrr.2023.100623 ( 2024 ). OpenUrl CrossRef ↵ Gay , C. M. et al. Patterns of transcription factor programs and immune pathway activation define four major subtypes of SCLC with distinct therapeutic vulnerabilities . Cancer Cell 39 , 346 – 360 e347 , doi: 10.1016/j.ccell.2020.12.014 ( 2021 ). OpenUrl CrossRef PubMed ↵ Rudin , C. M. et al. Molecular subtypes of small cell lung cancer: a synthesis of human and mouse model data . Nat Rev Cancer 19 , 289 – 297 , doi: 10.1038/s41568-019-0133-9 ( 2019 ). OpenUrl CrossRef PubMed ↵ Baine , M. K. et al. POU2F3 in SCLC: Clinicopathologic and Genomic Analysis With a Focus on Its Diagnostic Utility in Neuroendocrine-Low SCLC . J Thorac Oncol 17 , 1109 – 1121 , doi: 10.1016/j.jtho.2022.06.004 ( 2022 ). OpenUrl CrossRef PubMed ↵ Baine , M. K. et al. SCLC Subtypes Defined by ASCL1, NEUROD1, POU2F3, and YAP1: A Comprehensive Immunohistochemical and Histopathologic Characterization . J Thorac Oncol 15 , 1823 – 1835 , doi: 10.1016/j.jtho.2020.09.009 ( 2020 ). OpenUrl CrossRef PubMed ↵ Ireland , A. S. et al. MYC Drives Temporal Evolution of Small Cell Lung Cancer Subtypes by Reprogramming Neuroendocrine Fate . Cancer Cell 38 , 60 – 78 e12 , doi: 10.1016/j.ccell.2020.05.001 ( 2020 ). OpenUrl CrossRef PubMed ↵ Duplaquet , L. et al. KDM6A epigenetically regulates subtype plasticity in small cell lung cancer . Nat Cell Biol , doi: 10.1038/s41556-023-01210-z ( 2023 ). OpenUrl CrossRef PubMed ↵ Blackledge , N. P. & Klose , R. J . The molecular principles of gene regulation by Polycomb repressive complexes . Nat Rev Mol Cell Biol 22 , 815 – 833 , doi: 10.1038/s41580-021-00398-y ( 2021 ). OpenUrl CrossRef ↵ Kim , K. H. & Roberts , C. W . Targeting EZH2 in cancer . Nat Med 22 , 128 – 134 , doi: 10.1038/nm.4036 ( 2016 ). OpenUrl CrossRef PubMed ↵ Byers , L. A. et al. Proteomic profiling identifies dysregulated pathways in small cell lung cancer and novel therapeutic targets including PARP1 . Cancer Discov 2 , 798 – 811 , doi: 10.1158/2159-8290.CD-12-0112 ( 2012 ). OpenUrl Abstract / FREE Full Text ↵ Hong , D. et al. Plasticity in the Absence of NOTCH Uncovers a RUNX2-Dependent Pathway in Small Cell Lung Cancer . Cancer Res , doi: 10.1158/0008-5472.CAN-21-1991 ( 2021 ). OpenUrl Abstract / FREE Full Text ↵ Oser , M. G. et al. The KDM5A/RBP2 histone demethylase represses NOTCH signaling to sustain neuroendocrine differentiation and promote small cell lung cancer tumorigenesis . Genes Dev 33 , 1718 – 1738 , doi: 10.1101/gad.328336.119 ( 2019 ). OpenUrl Abstract / FREE Full Text ↵ Li , Y. et al. Aurora A kinase inhibition induces accumulation of SCLC tumor cells in mitosis with restored interferon signaling to increase response to PD-L1 . Cell Rep Med 4 , 101282 , doi: 10.1016/j.xcrm.2023.101282 ( 2023 ). OpenUrl CrossRef PubMed ↵ Oser , M. G. , MacPherson , D. , Oliver , T. G. , Sage , J. & Park , K. S . Genetically-engineered mouse models of small cell lung cancer: the next generation . Oncogene , doi: 10.1038/s41388-023-02929-7 ( 2024 ). OpenUrl CrossRef ↵ Ezhkova , E. et al. EZH1 and EZH2 cogovern histone H3K27 trimethylation and are essential for hair follicle homeostasis and wound repair . Genes Dev 25 , 485 – 498 , doi: 10.1101/gad.2019811 ( 2011 ). OpenUrl Abstract / FREE Full Text ↵ Poirier , J. T. et al. DNA methylation in small cell lung cancer defines distinct disease subtypes and correlates with high expression of EZH2 . Oncogene 34 , 5869 – 5878 , doi: 10.1038/onc.2015.38 ( 2015 ). OpenUrl CrossRef PubMed ↵ Schaffer , B. E. et al. Loss of p130 accelerates tumor development in a mouse model for human small-cell lung carcinoma . Cancer Res 70 , 3877 – 3883 , doi: 10.1158/0008-5472.CAN-09-4228 ( 2010 ). OpenUrl Abstract / FREE Full Text ↵ Liberzon , A. et al. The Molecular Signatures Database (MSigDB) hallmark gene set collection . Cell Syst 1 , 417 – 425 , doi: 10.1016/j.cels.2015.12.004 ( 2015 ). OpenUrl CrossRef PubMed ↵ Consortium , E. P . An integrated encyclopedia of DNA elements in the human genome . Nature 489 , 57 – 74 , doi: 10.1038/nature11247 ( 2012 ). OpenUrl CrossRef PubMed Web of Science ↵ Lachmann , A. et al. ChEA: transcription factor regulation inferred from integrating genome-wide ChIP-X experiments . Bioinformatics 26 , 2438 – 2444 , doi: 10.1093/bioinformatics/btq466 ( 2010 ). OpenUrl CrossRef PubMed Web of Science ↵ Shim , H. S. et al. Unique Genetic and Survival Characteristics of Invasive Mucinous Adenocarcinoma of the Lung . J Thorac Oncol 10 , 1156 – 1162 , doi: 10.1097/JTO.0000000000000579 ( 2015 ). OpenUrl CrossRef PubMed ↵ Guo , M. et al. Gene signature driving invasive mucinous adenocarcinoma of the lung . EMBO Mol Med 9 , 462 – 481 , doi: 10.15252/emmm.201606711 ( 2017 ). OpenUrl Abstract / FREE Full Text Snyder , E. L. et al. Nkx2-1 represses a latent gastric differentiation program in lung adenocarcinoma . Mol Cell 50 , 185 – 199 , doi: 10.1016/j.molcel.2013.02.018 ( 2013 ). OpenUrl CrossRef PubMed Web of Science ↵ Di Federico , A. et al. Lung adenocarcinomas with mucinous histology: clinical, genomic, and immune microenvironment characterization and outcomes to immunotherapy-based treatments and KRAS(G12C) inhibitors . Ann Oncol , doi: 10.1016/j.annonc.2024.11.014 ( 2024 ). OpenUrl CrossRef ↵ Tickle , T. T. , Itay; Georgescu, Christophe; Brown, Maxwell; Haas, Brian. inferCNV of the Trinity CTAT Project . https://github.com/broadinstitute/inferCNV Klarman Cell Observatory, Broad Institute of MIT and Harvard , Cambridge, MA, USA ( 2019 ). ↵ Duan , Z. & Luo , Y . Targeting macrophages in cancer immunotherapy . Signal Transduct Target Ther 6 , 127 , doi: 10.1038/s41392-021-00506-6 ( 2021 ). OpenUrl CrossRef ↵ Pittet , M. J. , Michielin , O. & Migliorini , D . Clinical relevance of tumour-associated macrophages . Nat Rev Clin Oncol 19 , 402 – 421 , doi: 10.1038/s41571-022-00620-6 ( 2022 ). OpenUrl CrossRef PubMed ↵ Browaeys , R. , Saelens , W. & Saeys , Y . NicheNet: modeling intercellular communication by linking ligands to target genes . Nat Methods 17 , 159 – 162 , doi: 10.1038/s41592-019-0667-5 ( 2020 ). OpenUrl CrossRef PubMed ↵ Calbo , J. et al. A functional role for tumor cell heterogeneity in a mouse model of small cell lung cancer . Cancer Cell 19 , 244 – 256 , doi: 10.1016/j.ccr.2010.12.021 ( 2011 ). OpenUrl CrossRef PubMed Web of Science ↵ Jiang , J. et al. Translational and Therapeutic Evaluation of RAS-GTP Inhibition by RMC-6236 in RAS-Driven Cancers . Cancer Discov 14 , 994 – 1017 , doi: 10.1158/2159-8290.CD-24-0027 ( 2024 ). OpenUrl CrossRef PubMed ↵ Jadhav , U. et al. Acquired Tissue-Specific Promoter Bivalency Is a Basis for PRC2 Necessity in Adult Cells . Cell 165 , 1389 – 1400 , doi: 10.1016/j.cell.2016.04.031 ( 2016 ). OpenUrl CrossRef PubMed ↵ Mahadevan , N. R. et al. Intrinsic immunogenicity of small cell lung carcinoma revealed by its cellular plasticity . Cancer Discov , doi: 10.1158/2159-8290.CD-20-0913 ( 2021 ). OpenUrl Abstract / FREE Full Text ↵ Burr , M. L. et al. An Evolutionarily Conserved Function of Polycomb Silences the MHC Class I Antigen Presentation Pathway and Enables Immune Evasion in Cancer . Cancer Cell 36 , 385 – 401 e388 , doi: 10.1016/j.ccell.2019.08.008 ( 2019 ). OpenUrl CrossRef PubMed ↵ Politi , K. et al. Lung adenocarcinomas induced in mice by mutant EGF receptors found in human lung cancers respond to a tyrosine kinase inhibitor or to down-regulation of the receptors . Genes Dev 20 , 1496 – 1510 , doi: 10.1101/gad.1417406 ( 2006 ). OpenUrl Abstract / FREE Full Text ↵ Foggetti , G. et al. Genetic Determinants of EGFR-Driven Lung Cancer Growth and Therapeutic Response In Vivo . Cancer Discov 11 , 1736 – 1753 , doi: 10.1158/2159-8290.CD-20-1385 ( 2021 ). OpenUrl Abstract / FREE Full Text ↵ Mazutis ; Triparna Sen ; Ronan Chaligne ; Helena Yu ; Dana Pe’er ; Charles Rudin , J. M. C. A. Q.-V. A. S. P. M. O. C. T. X. I. M. J. E. N. S. J. C. T. N. L. Abstract 1167: Single-cell transcriptomic profiling of SCLC transformation reveals increased intratumoral diversity of variant and non-neuroendocrine subtypes AACR Annual Meeting 2023 Oral Abstract ( 2023 ). ↵ Mollaoglu , G. et al. MYC Drives Progression of Small Cell Lung Cancer to a Variant Neuroendocrine Subtype with Vulnerability to Aurora Kinase Inhibition . Cancer Cell 31 , 270 – 285 , doi: 10.1016/j.ccell.2016.12.005 ( 2017 ). OpenUrl CrossRef PubMed ↵ El Zarif , T. , et al. Detecting Small Cell Transformation in Patients with Advanced EGFR Mutant Lung Adenocarcinoma through Epigenomic cfDNA Profiling . Clin Cancer Res 30 , 3798 – 3811 , doi: 10.1158/1078-0432.CCR-24-0466 ( 2024 ). OpenUrl CrossRef PubMed ↵ Bracken , A. P. et al. EZH2 is downstream of the pRB-E2F pathway, essential for proliferation and amplified in cancer . EMBO J 22 , 5323 – 5335 , doi: 10.1093/emboj/cdg542 ( 2003 ). OpenUrl Abstract / FREE Full Text ↵ Ahn , M. J. et al. Tarlatamab for Patients with Previously Treated Small-Cell Lung Cancer . N Engl J Med 389 , 2063 – 2075 , doi: 10.1056/NEJMoa2307980 ( 2023 ). OpenUrl CrossRef ↵ Paz-Ares , L. et al. Tarlatamab, a First-in-Class DLL3-Targeted Bispecific T-Cell Engager, in Recurrent Small-Cell Lung Cancer: An Open-Label, Phase I Study . J Clin Oncol 41 , 2893 – 2903 , doi: 10.1200/JCO.22.02823 ( 2023 ). OpenUrl CrossRef ↵ Serresi , M. et al. Polycomb Repressive Complex 2 Is a Barrier to KRAS-Driven Inflammation and Epithelial-Mesenchymal Transition in Non-Small-Cell Lung Cancer . Cancer Cell 29 , 17 – 31 , doi: 10.1016/j.ccell.2015.12.006 ( 2016 ). OpenUrl CrossRef PubMed ↵ Zhang , H. et al. Lkb1 inactivation drives lung cancer lineage switching governed by Polycomb Repressive Complex 2 . Nat Commun 8 , 14922 , doi: 10.1038/ncomms14922 ( 2017 ). OpenUrl CrossRef PubMed ↵ Zorn , A. M. & Wells , J. M . Vertebrate endoderm development and organ formation . Annu Rev Cell Dev Biol 25 , 221 – 251 , doi: 10.1146/annurev.cellbio.042308.113344 ( 2009 ). OpenUrl CrossRef PubMed Web of Science ↵ Borromeo , M. D. et al. ASCL1 and NEUROD1 Reveal Heterogeneity in Pulmonary Neuroendocrine Tumors and Regulate Distinct Genetic Programs . Cell Rep 16 , 1259 – 1272 , doi: 10.1016/j.celrep.2016.06.081 ( 2016 ). OpenUrl CrossRef PubMed ↵ Naya , F. J. , Stellrecht , C. M. & Tsai , M. J . Tissue-specific regulation of the insulin gene by a novel basic helix-loop-helix transcription factor . Genes Dev 9 , 1009 – 1019 , doi: 10.1101/gad.9.8.1009 ( 1995 ). OpenUrl Abstract / FREE Full Text Gu , C. et al. Pancreatic beta cells require NeuroD to achieve and maintain functional maturity . Cell Metab 11 , 298 – 310 , doi: 10.1016/j.cmet.2010.03.006 ( 2010 ). OpenUrl CrossRef PubMed Web of Science ↵ Bohuslavova , R. et al. NEUROD1 reinforces endocrine cell fate acquisition in pancreatic development . Nat Commun 14 , 5554 , doi: 10.1038/s41467-023-41306-6 ( 2023 ). OpenUrl CrossRef PubMed ↵ Jennings , R. E. et al. Development of the human pancreas from foregut to endocrine commitment . Diabetes 62 , 3514 – 3522 , doi: 10.2337/db12-1479 ( 2013 ). OpenUrl Abstract / FREE Full Text ↵ Romero , R. et al. The neuroendocrine transition in prostate cancer is dynamic and dependent on ASCL1 . Nat Cancer 5 , 1641 – 1659 , doi: 10.1038/s43018-024-00838-6 ( 2024 ). OpenUrl CrossRef ↵ Rudin , C. M. et al. Clinical Benefit From Immunotherapy in Patients With SCLC Is Associated With Tumor Capacity for Antigen Presentation . J Thorac Oncol 18 , 1222 – 1232 , doi: 10.1016/j.jtho.2023.05.008 ( 2023 ). OpenUrl CrossRef PubMed ↵ Yang , Y. & Fan , S . Small cell lung cancer transformations from non-small cell lung cancer: Biological mechanism and clinical relevance . Chin Med J Pulm Crit Care Med 2 , 42 – 47 , doi: 10.1016/j.pccm.2023.10.005 ( 2024 ). OpenUrl CrossRef PubMed ↵ Wagle , M. C. et al. A transcriptional MAPK Pathway Activity Score (MPAS) is a clinically relevant biomarker in multiple cancer types . NPJ Precis Oncol 2 , 7 , doi: 10.1038/s41698-018-0051-4 ( 2018 ). OpenUrl CrossRef PubMed ↵ Cerami , E. et al. The cBio cancer genomics portal: an open platform for exploring multidimensional cancer genomics data . Cancer Discov 2 , 401 – 404 , doi: 10.1158/2159-8290.CD-12-0095 ( 2012 ). OpenUrl Abstract / FREE Full Text ↵ DuPage , M. , Dooley , A. L. & Jacks , T . Conditional mouse lung cancer models using adenoviral or lentiviral delivery of Cre recombinase . Nat Protoc 4 , 1064 – 1072 , doi: 10.1038/nprot.2009.95 ( 2009 ). OpenUrl CrossRef PubMed ↵ Subramanian , A. et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles . Proc Natl Acad Sci U S A 102 , 15545 – 15550 , doi: 10.1073/pnas.0506580102 ( 2005 ). OpenUrl Abstract / FREE Full Text ↵ Kanehisa , M. , Furumichi , M. , Sato , Y. , Matsuura , Y. & Ishiguro-Watanabe , M . KEGG: biological systems database as a model of the real world . Nucleic Acids Res 53 , D672 – D677 , doi: 10.1093/nar/gkae909 ( 2025 ). OpenUrl CrossRef ↵ Su , A. I. et al. A gene atlas of the mouse and human protein-encoding transcriptomes . Proc Natl Acad Sci U S A 101 , 6062 – 6067 , doi: 10.1073/pnas.0400782101 ( 2004 ). OpenUrl Abstract / FREE Full Text ↵ Gene Ontology, C ., et al. The Gene Ontology knowledgebase in 2023 . Genetics 224 , doi: 10.1093/genetics/iyad031 ( 2023 ). OpenUrl CrossRef PubMed ↵ Chen , E. Y. et al. Enrichr: interactive and collaborative HTML5 gene list enrichment analysis tool . BMC Bioinformatics 14 , 128 , doi: 10.1186/1471-2105-14-128 ( 2013 ). OpenUrl CrossRef PubMed ↵ Taing , L. et al. CHIPS: A Snakemake pipeline for quality control and reproducible processing of chromatin profiling data [version 1; peer review: 1 not approved] . F1000Research 10 , doi: 10.12688/f1000research.52878.1 ( 2021 ). OpenUrl CrossRef ↵ Qiu , X. et al. CoBRA: Containerized Bioinformatics Workflow for Reproducible ChIP/ATAC-seq Analysis . Genomics, proteomics & bioinformatics 19 , 652 – 661 , doi: 10.1016/j.gpb.2020.11.007 ( 2021 ). OpenUrl CrossRef PubMed ↵ Li , H. & Durbin , R . Fast and accurate short read alignment with Burrows-Wheeler transform . Bioinformatics 25 , 1754 – 1760 , doi: 10.1093/bioinformatics/btp324 ( 2009 ). OpenUrl CrossRef PubMed Web of Science ↵ Zhang , Y. et al. Model-based analysis of ChIP-Seq (MACS) . Genome Biol 9 , R137 – R137 , doi: 10.1186/gb-2008-9-9-r137 ( 2008 ). OpenUrl CrossRef PubMed ↵ Robinson , J. T. et al. Integrative genomics viewer . Nat Biotechnol 29 , 24 – 26 , doi: 10.1038/nbt.1754 ( 2011 ). OpenUrl CrossRef PubMed Web of Science ↵ Ross-Innes , C. S. et al. Differential oestrogen receptor binding is associated with clinical outcome in breast cancer . Nature 481 , 389 – 393 , doi: 10.1038/nature10730 ( 2012 ). OpenUrl CrossRef PubMed Web of Science ↵ Heinz , S. et al. Simple combinations of lineage-determining transcription factors prime cis-regulatory elements required for macrophage and B cell identities . Molecular cell 38 , 576 – 589 , doi: 10.1016/j.molcel.2010.05.004 ( 2010 ). OpenUrl CrossRef PubMed Web of Science ↵ Ramirez , F. et al. deepTools2: a next generation web server for deep-sequencing data analysis . Nucleic Acids Res 44 , W160 – 165 , doi: 10.1093/nar/gkw257 ( 2016 ). OpenUrl CrossRef PubMed ↵ Li , H. & Durbin , R . Fast and accurate long-read alignment with Burrows-Wheeler transform . Bioinformatics 26 , 589 – 595 , doi: 10.1093/bioinformatics/btp698 ( 2010 ). OpenUrl CrossRef PubMed Web of Science ↵ Neph , S. et al. BEDOPS: high-performance genomic feature operations . Bioinformatics 28 , 1919 – 1920 , doi: 10.1093/bioinformatics/bts277 ( 2012 ). OpenUrl CrossRef PubMed Web of Science ↵ Duplaquet , L. et al. KDM6A epigenetically regulates subtype plasticity in small cell lung cancer . Nat Cell Biol 25 , 1346 – 1358 , doi: 10.1038/s41556-023-01210-z ( 2023 ). OpenUrl CrossRef PubMed ↵ Platt , R. J. et al. CRISPR-Cas9 knockin mice for genome editing and cancer modeling . Cell 159 , 440 – 455 , doi: 10.1016/j.cell.2014.09.014 ( 2014 ). OpenUrl CrossRef PubMed Web of Science ↵ Hao , Y. et al. Integrated analysis of multimodal single-cell data . Cell 184 , 3573 – 3587 e3529 , doi: 10.1016/j.cell.2021.04.048 ( 2021 ). OpenUrl CrossRef PubMed ↵ Aran , D. et al. Reference-based analysis of lung single-cell sequencing reveals a transitional profibrotic macrophage . Nat Immunol 20 , 163 – 172 , doi: 10.1038/s41590-018-0276-y ( 2019 ). OpenUrl CrossRef PubMed ↵ Heng , T. S. , Painter , M. W. & Immunological Genome Project, C . The Immunological Genome Project: networks of gene expression in immune cells . Nat Immunol 9 , 1091 – 1094 , doi: 10.1038/ni1008-1091 ( 2008 ). OpenUrl CrossRef PubMed Web of Science ↵ Benayoun , B. A. et al. Remodeling of epigenome and transcriptome landscapes with aging in mice reveals widespread induction of inflammatory responses . Genome Res 29 , 697 – 709 , doi: 10.1101/gr.240093.118 ( 2019 ). OpenUrl Abstract / FREE Full Text ↵ Nei , M. & Kumar , S . Molecular evolution and phylogenetics . ( Oxford University Press , 2000 ). ↵ Tamura , K. , Stecher , G. & Kumar , S . MEGA11: Molecular Evolutionary Genetics Analysis Version 11 . Mol Biol Evol 38 , 3022 – 3027 , doi: 10.1093/molbev/msab120 ( 2021 ). OpenUrl CrossRef PubMed ↵ McLean , C. Y. et al. GREAT improves functional interpretation of cis-regulatory regions . Nat Biotechnol 28 , 495 – 501 , doi: 10.1038/nbt.1630 ( 2010 ). OpenUrl CrossRef PubMed Web of Science View the discussion thread. Back to top Previous Next Posted July 10, 2025. Download PDF Supplementary Material Email Thank you for your interest in spreading the word about bioRxiv. NOTE: Your email address is requested solely to identify you as the sender of this article. Your Email * Your Name * Send To * Enter multiple addresses on separate lines or separate them with commas. You are going to email the following EED Maintains the Small Cell Lung Cancer Neuroendocrine Phenotype and Drives Lung Cancer Histological Transformation Message Subject (Your Name) has forwarded a page to you from bioRxiv Message Body (Your Name) thought you would like to see this page from the bioRxiv website. Your Personal Message CAPTCHA This question is for testing whether or not you are a human visitor and to prevent automated spam submissions. Share EED Maintains the Small Cell Lung Cancer Neuroendocrine Phenotype and Drives Lung Cancer Histological Transformation Yixiang Li , Yasmin N. Laimon , Hyeonseo Cho , Marina Vivero , Gabriel Roberti De Oliveira , Andrew Delcea , Varunika Savla , Yuting Chen , Yavuz T. Durmaz , Xintao Qiu , Shweta Kukreja , Rong Li , Talal El Zarif , Wesley Lu , McKayla Van Orden , Jacob E. Berchuck , Roderick T. Bronson , Shuqiang Li , Hongbin Ji , Katerina Politi , Matthew L. Freedman , Henry W. Long , Sabina Signoretti , Matthew G. Oser bioRxiv 2025.07.07.663486; doi: https://doi.org/10.1101/2025.07.07.663486 Share This Article: Copy Citation Tools EED Maintains the Small Cell Lung Cancer Neuroendocrine Phenotype and Drives Lung Cancer Histological Transformation Yixiang Li , Yasmin N. Laimon , Hyeonseo Cho , Marina Vivero , Gabriel Roberti De Oliveira , Andrew Delcea , Varunika Savla , Yuting Chen , Yavuz T. Durmaz , Xintao Qiu , Shweta Kukreja , Rong Li , Talal El Zarif , Wesley Lu , McKayla Van Orden , Jacob E. Berchuck , Roderick T. Bronson , Shuqiang Li , Hongbin Ji , Katerina Politi , Matthew L. Freedman , Henry W. Long , Sabina Signoretti , Matthew G. Oser bioRxiv 2025.07.07.663486; doi: https://doi.org/10.1101/2025.07.07.663486 Citation Manager Formats BibTeX Bookends EasyBib EndNote (tagged) EndNote 8 (xml) Medlars Mendeley Papers RefWorks Tagged Ref Manager RIS Zotero Tweet Widget Facebook Like Google Plus One Subject Area Cancer Biology Subject Areas All Articles Animal Behavior and Cognition (7635) Biochemistry (17691) Bioengineering (13892) Bioinformatics (41936) Biophysics (21452) Cancer Biology (18588) Cell Biology (25504) Clinical Trials (138) Developmental Biology (13378) Ecology (19899) Epidemiology (2067) Evolutionary Biology (24320) Genetics (15609) Genomics (22506) Immunology (17736) Microbiology (40394) Molecular Biology (17181) Neuroscience (88605) Paleontology (666) Pathology (2832) Pharmacology and Toxicology (4824) Physiology (7641) Plant Biology (15153) Scientific Communication and Education (2045) Synthetic Biology (4294) Systems Biology (9825) Zoology (2271)