Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 Mar 1.
Published in final edited form as: J Thorac Oncol. 2022 Dec 5;18(3):324–338. doi: 10.1016/j.jtho.2022.11.019

Pten loss expands the histopathological diversity and lineage plasticity of lung cancers initiated by Rb1/Trp53 deletion

Letian Zhang 1,2, Congrong Liu 3, Bo Zhang 3, Jie Zheng 3, Prashant Singh 4, Wiam Bshara 5, Jianmin Wang 6, Eduardo Cortes Gomez 6, Xiaojing Zhang 1, Yanqing Wang 1, Xiang Zhu 3,7, David W Goodrich 1,7
PMCID: PMC9974779  NIHMSID: NIHMS1861659  PMID: 36473627

Abstract

High grade neuroendocrine (NE) tumors of the lung, like small cell lung cancer (SCLC), are recalcitrant cancers for which more effective systemic therapies are needed. Despite their histopathological and molecular heterogeneity, they are generally treated as a single disease entity with similar chemotherapy regimens. While marked clinical responses can be observed, they are short lived. Inter- and intra-tumoral heterogeneity is considered a confounding factor in these unsatisfactory clinical outcomes, yet the origin of this heterogeneity and its impact on therapeutic responses is not well understood. Here new genetically engineered mouse models are employed to test the effects of Pten loss on the development of lung tumors initiated by Rb1 and Trp53 tumor suppressor gene deletion. Complete Pten loss drives more rapid tumor development with a greater diversity of tumor histopathology ranging from adenocarcinoma to SCLC. Pten loss also drives transcriptional heterogeneity as marked lineage plasticity is observed within histopathological subtypes. Spatial profiling indicates transcriptional heterogeneity exists both within and between tumor foci with limited spatial intermixing of transcriptional clusters, implying that the growth environment influences gene expression. These results identify Pten loss as clinically relevant genetic alteration driving the molecular and histopatholgical heterogeneity of NE lung tumors initiated by Rb1/Trp53 mutations.

Keywords: small cell lung cancer, tumor heterogeneity, genetically engineered mouse models, tumor suppressor genes

Introduction

Small-cell lung carcinoma (SCLC) is a highly lethal and metastatic high-grade neuroendocrine (NE) carcinoma with median patient survival ranging from 8–20 months1–3. SCLC has an aggressive disease course such that at the time of diagnosis most patients already present with metastatic disease. Five-year survival rates are poor at less than 7%. The standard of care for treating SCLC remains chemotherapy, with potential use of immunotherapy more recently considered. While excellent clinical responses can be observed, therapeutic resistance develops rapidly resulting in median survival duration of about one year. Based on these disappointing therapeutic outcomes, SCLC has been classified as a recalcitrant disease1. Since SCLC accounts for about 15% of all lung cancers, it remains a significant clinical problem.

SCLC is treated as a single disease entity, yet this disease exhibits marked intra-tumoral and inter-tumoral heterogeneity that may contribute to poor therapeutic outcomes4, 5. Histopathological assessment of SCLC reveals highly cellular tumors composed of undifferentiated small blue cells with scant cytoplasm and granular chromatin6, 7. Yet a variety of additional cellular characteristics may be observed like large cells, pleiomorphic cells, or giant cells. Tumors commonly grow as diffuse sheets but can also exhibit palisading, streams, ribbons, organoid nesting, and rosettes8, 9. SCLC is often observed in conjunction with non-small cell lung carcinoma (NSCLC) elements like adenocarcinoma or large cell neuroendocrine carcinoma9, 10. Histological transformation from lung adenocarcinoma to an NE cancer resembling SCLC has also been reported, typically in association with acquired resistance to molecularly targeted therapies like EGFR tyrosine kinase inhibitors11–14. In addition to this histopathological heterogeneity, transcriptional profiling has distinguished multiple molecular subtypes based on the expression of lineage specifying transcription factors ASCL1, NEUROD1, POU2F3, and YAP115. These molecular subtypes can exhibit differential sensitivity to cancer therapies16. Accumulating data thus suggests clinical outcomes may be influenced by molecular and/or histologic heterogeneity15, 17.

The mutational landscape of SCLC is dominated by inactivating alterations in the RB1 and TP53 tumor suppressor genes that occur in most cases18–21, although not all RB1/TP53 mutant lung tumors develop into SCLC. Other genetic alterations like those in the PIK3A/PTEN pathway are also recurrent in SCLC. Yet, none of these genetic alterations clearly associate with molecular or histologic subtypes. Thus, it remains incompletely understood how molecular and histopathological heterogeneity arises in SCLC. As SCLC is not often resected in the clinic due to its highly metastatic nature, genetically engineered mouse models have become important tools to address this knowledge gap22, 23. Rb1/Trp53 deletion in the mouse lung is sufficient to initiate development of SCLC after long latency24. During this latent period, murine SCLCs can acquire other mutations like loss of function Pten mutations25, 26. Deletion of Rb1/Trp53 in pulmonary neuroendocrine cells causes highly penetrant SCLC indicating these cells can serve as the cell of SCLC origin27, 28. However, SCLC can also arise in mice when Rb1/Trp53 are deleted in other cell types28–31, including SPC expressing AT2 alveolar cells. How the genetic background of different cells of cancer origin influence the histologic and molecular heterogeneity of lung cancers is not well understood. Here we address this issue by performing detailed histopathological analysis and gene expression profiling of tumors developing in genetically engineered mouse models of lung cancer initiated by defined genetic alterations in specified cells of cancer origin.

Material and Methods

Mice and Adenoviral Infection

Rb1flox, Trp53flox, and Ptenflox mouse alleles have been described previously24, 32. Genotyping was performed by PCR analysis of genomic DNA extracted from tail clips as described previously33, 34. Experimental mice were on a mixed genetic background (C57BL/6:129/Sv). Mice were monitored daily, euthanized when moribund, and necropsied to verify diagnosis and collect tissue. Survival analysis by the Kaplan-Meier method was done with GraphPad Prism software. Ad-CMV-Cre, and Ad-SPC-Cre were described previously28 and purchased from the University of Iowa viral vector core. 108 pfu Adeno-Cre virus was administrated per mouse using intratracheal injection35. All animal experiments comply with the Guide for the Care and Use of Laboratory Animals and has been approved by the Institutional Animal Care and Use Committee at Roswell Park.

Histology and Immunohistochemistry (IHC)

Lung tissue was fixed in phosphate-buffered 4% paraformaldehyde, embedded in paraffin, and serially sectioned at 5-μm thickness. Sections were H&E stained for assessment of histopathology (C.L., B.Z., J.Z., X.J., W.B. reviewing pathologists). Primary antibodies and antibody dilutions used for IHC were: Synaptophysin (Thermofisher #PA5–16417; 1:600), TTF-1 (Invitrogen #MA5–13961; 1:600;), Ki67 (Leica #NCL-Ki67p; 1:1000;), and SPC (Sigma-Aldrich # AB3786; 1:2000). Immunostaining was developed using Diaminobenzidine (DAB) (Dako, K3468) followed by hematoxylin counterstaining.

Pathology

The diagnosis of histology subtypes of mice was based on World Health Organization classification of tumors of the lung. SCLCs are malignant epithelial tumors expressing neuroendocrine (NE) markers of small cells, usually round, oval to spindle in shape with scant cytoplasm. Nucleoli are absent and inconspicuous with finely granular nuclear chromatin, frequently with high mitotic rate. Tumor cells grow in sheet-like or nest-like patten, frequently with necrosis. LCNEC is high-grade non-small cell carcinoma expressing NE markers of large cells with neuroendocrine morphology, such as organoid nesting and rosette pattern. Cells are moderate to abundant cytoplasm with distinct borders. Some tumors have fine nuclear chromatin with nucleoli features analogous to SCLC. Adenocarcinomas (ADs) usually grow in mixed patterns including lepidic, acinar, solid, papillary and micropapillary. Adenocarcinoma with NE differentiation (AD-NE) is defined as tumors with adenocarcinoma morphology but NE markers expression, indicating NE differentiation. Atypical papillary bronchiolar proliferation (APBP) is bronchiolar epithelium proliferation. The most significant feature is papillary proliferation with a central fibrovascular core and mildly atypical columnar epithelial cells36.

A central location for tumors is defined as those that grow in and/or around major central airways. A peripheral location is defined as tumors that grow in and/or around terminal airways in the peripheral lung. A diffuse pattern is defined as tumors significantly covering most of the lung. Tumor burden was determined based on the percentage of lung area occupied by tumors as measured using ImageJ software. Histological quantification was performed by quantifying the frequency of cells that were positive for specified antigens and staining intensity scoring from 0 (negative) to 3 (very dark staining).

RNA Sequencing

The sequencing libraries were prepared with the RNA HyperPrep Kit with RiboErase (HMR) kit (Roche Sequencing Solutions), from 500ng total RNA, following manufacturer’s instructions. The first step depletes rRNA from total RNA. After ribosomal depletion, the remaining RNA is DNase digested to remove gDNA contamination. Samples were then purified, fragmented, and primed for cDNA synthesis. Fragmented RNA was then reverse transcribed into first strand cDNA using random primers. The next step removed the RNA template and synthesized a replacement strand, incorporating dUTP in place of dTTP to generate ds cDNA. Pure Beads (KAPA BIOSYSTEMS) were used to separate the ds cDNA from the second strand reaction mix resulting in blunt-ended cDNA. A single ‘A’ nucleotide is then added to the 3’ ends of the blunt fragments. Multiple indexing adapters, containing a single ‘T’ nucleotide on the 3’ end of the adapter, were ligated to the ends of the ds cDNA, preparing them for hybridization onto a flow cell. Adapter ligated libraries were amplified by PCR, purified using Pure Beads, and validated for appropriate size on a 4200 TapeStation D1000 Screentape (Agilent Technologies, Inc.). The DNA libraries were quantitated using KAPA Biosystems qPCR kit, and were pooled together in an equimolar fashion. Each pool was denatured and diluted to 350 pM with 1% PhiX control library added. The resulting pool was then loaded into a 200 cycle NovaSeq Reagent cartridge for 100 cycle paired end sequencing using a NovaSeq6000 following the manufacturer’s recommended protocol (Illumina Inc.).

RNA Sequencing Analysis

Paired-end raw sequencing reads passing quality filters from Illumina RTA were first pre-processed using FASTQC (v0.10.1) for sequencing base quality control37. The reads were mapped to the mm10 mouse reference genome and corresponding RefSeq38 gene annotation database using splicing aware tools Bowtie (v1.0.1)39 and TopHat (v2.1.1)40 allowing a maximum of 1 mismatch per read. A second pass QC was done using alignment output with RSeQC (v2.6.3)41 in order to examine the abundance of genomic features, splicing junction saturation, and gene-body coverage. Gene expression was quantified using featureCounts from the Subread package (v1.6.0)42 with the --fracOverlap 1 option. Differential expression analyses were performed using DESeq2 (v1.18.1)43. Output results were visually presented using the pheatmap (v1.0.8)44 R package.

Clustering Analysis

Top 300 IQR (interquantile-range) regularized log2 transformed expression genes were used to run Hierarchical Clustering Analysis to determine prominent clustering profiles. The IQR statistic was used to robustly detect highly variable genes. The data was fitted with the hclust function from the stats native R (v4.1.1) package, using “Complete” clustering method paired with Euclidean distance. Clustering groups were selected by cutting the 4 main branches resulting from the hierarchical clustering tree. The clusters were manually curated and given the following labels based on their gene profiles: “cl1-Adeno” (1), “cl2-LungClub” (3), “cl3-Dediff” (5) and “other” (4). Only the first three groups were considered for downstream analyses. Pairwise group differential expression analysis was carried out, using DESeq2, to determine over-expressed sets of genes that characterize each clustering group. GSEA [11] was then used to further examine the association of the cluster groups with canonical biological pathways using C2cp database from msigDB (v7.4)45.

Spatial Transcriptomic Profiling

Spatial profiling was performed using 10x Visium Spatial for FFPE Gene Expression Kit (Spatial 3’ v1, 10X Genomics Inc.). Briefly, tissue embedded paraffin blocks were sectioned and trimmed to fit within the four capture areas on the Visium Spatial slides. Deparaffinization and H&E staining was performed, followed by imaging of each tissue sample. The RNAs within the tissue were then hybridized to mouse whole transcriptome probe panel and hybridized probes were captured on the Visium slides. Captured probe products were extended with the addition of an UMI, Spatial Barcode and partial Read 1, thus synthesizing cDNA to be used for gene expression library construction. FFPE gene expression libraries for each sample were produced with enzymatic fragmentation, end-repair, a-tailing, adapter ligation, and PCR to add Illumina compatible sequencing adapters. The resulting libraries were evaluated on D1000 screentape using a TapeStation 4200 (Agilent Technologies) and quantitated using KAPA Biosystems qPCR quantitation kit for Illumina. They were then pooled, denatured, and diluted to 300pM with 1% PhiX control library added. The resulting pool was then loaded into the appropriate NovaSeq Reagent cartridge and sequenced on a NovaSeq6000 following the manufacturer’s recommended protocol (Illumina Inc.). Once sequencing was complete, tissue images taken after H&E staining were used to align the FFPE gene expression from the spatial barcodes unique to each location in the capture area during data analysis using 10x Genomics Space Ranger v 1.3.1 software.

The raw sequencing data, mapping results (BAM files) and quantification matrices were generated using cellranger software with mouse mm10 genome and GENCODE annotation database. Seurat single cell data analysis R package was used for spatial RNASeq data analysis. The gene counts matrices were normalized with SCTransform method and dimension reductions including principal component analysis (PCA), UMAP and tSNE were carried out the highly variable genes. Data clustering is identified using the shared nearest neighbor (SNN)-based clustering on the first 23 principal components. The neuroendocrine score (NE score) was calculated using AddModuleScore method with a list of NE marker genes.

Results

Pten loss accelerates lung cancer progression initiated by Rb1/Trp53 loss.

We generated mice with the Rb1flox/flox;Trp53flox/flox;Ptenflox/+ or Rb1flox/flox;Trp53flox/flox;Ptenflox/flox genotypes to assess effects of Pten loss on lung cancer progression initiated by Rb1/Trp53 double knockout (DKO)(Figure 1A). Tumorigenesis was initiated by intratracheal injection of Ad-CMV-Cre or Ad-SPC-Cre. Ad-CMV-Cre expresses Cre in a broad range of lung cell types while Ad-SPC-Cre restricts Cre mediated recombination to surfactant protein C (SPC) expressing AT2 cells26, 28. Homozygous deletion of Pten in DKO mice reduced mouse survival, relative to mice retaining one wild type Pten allele, when tumorigenesis was initiated by either Ad-CMV-Cre or Ad-SPC-Cre (Figure 1B). For Ad-CMV-Cre initiated mice, median survival was significantly different between DKO;Ptenflox/flox (119 days post-viral administration) and DKO;Ptenflox/+ mice (203 days, log rank P<0.01), consistent with published reports26, 28. Survival was longer in all Ad-SPC-Cre initiated cohorts, but the median survival for DKO;Ptenflox/flox mice (259 days) was still significantly shorter than for DKO;Ptenflox/+ mice (364 days, log rank P<0.01). Cre mediated deletion of floxed Rb1, Trp53 and Pten alleles was confirmed by PCR (Figure 1C). Thus complete Pten loss accelerates development of lethal lung cancers compared to loss of one Pten allele, both when initiated in a broad range of lung cell types or when initiated in SPC-expressing AT2 cells.

Figure 1. Pten loss accelerates lung cancer associated morbidity.

Figure 1.

Open in a new tab

A) Experimental scheme depicting administration of adenovirus designed to express Cre and delete indicated floxed alleles in various lung cell types, either unrestricted (CMV) or restricted to SPC expressing AT2 cells (SPC). B) Kaplan–Meier survival curves of mice in the indicated cohorts. Differences in the survival curves are statistically significant (log rank P<0.01). Cohort sample size ranges from 8–11 mice. C) Cre mediated deletion of indicated alleles was verified by PCR analysis of DNA extracted from the tail (unrecombined) or tumor (recombined). The image shows representative images of gel electrophoresis resolved PCR amplified DNA with expected bands indicated. Wild type mice and reactions lacking genomic DNA are used as controls. D) Tumor burden of mice at end stage in the indicated cohorts measured as the fractional area of lung tissue comprised of tumor. Each dot represents one mouse. Differences in cohorts are statistically significant (one-way ANOVA P<0.01). E) The incidence of metastasis is shown based on the fraction of mice in each cohort with detectable liver metastasis.

Primary tumor burden at disease endpoint was measured for each cohort based on the fractional area of lung tissue occupied by tumors (Figure 1D). Tumor burden was different in the four cohorts (one way ANOVA P<0.01) with greater average tumor burden in Ad-SPC-Cre initiated mice. Tumor burden at end stage was significantly greater for Ad-SPC-Cre initiated DKO;Ptenflox/flox mice than for Ad-CMV-Cre initiated DKO;Ptenflox/flox mice (t test P<0.01), for example. Given the lack of correlation between primary tumor burden and survival, we assessed mice for evidence of metastasis. All cohorts had evidence of metastatic dissemination from the lung to the liver (Figure 1E). However, the fraction of mice exhibiting detectable liver metastasis was higher for Ad-SPC-Cre initiated cohorts than for Ad-CMV-Cre initiated cohorts. The reduced survival of Ad-CMV-Cre initiated mice, therefore, was not explained by increased primary or metastatic tumor burden. Instead, primary and metastatic tumor burden correlated with survival time suggesting extended survival allowed more time for primary tumor growth and metastatic dissemination.

Pten loss diversifies lung cancer histopathology.

We performed detailed histopathological characterization of end stage tumors from mice in each of the cohorts to determine if differences in tumor phenotype might correlate with mouse survival. The criteria used for classification were based on standard definitions according to the WHO as well as published literature10, 22. Typically, mice contained multiple tumors distributed across all lobes of the lung, and multiple tumors within a given mouse often exhibited different histopathology. The two major tumor types observed in Ad-CMV-Cre initiated DKO;Ptenflox/flox mice, based on tumor area, were SCLC and adenocarcinoma with NE differentiation (AD-NE). AD-NE are defined as tumors with adenocarcinoma histology but that express NE markers like synaptophysin. Large cell neuroendocrine carcinoma (LCNEC) and adenocarcinoma (AD) were also observed (Figure 2A,B). Lung cancers in Ad-CMV-Cre initiated DKO;Ptenflox/+ mice were primarily SCLC and largely devoid of AD. Complete Pten loss, therefore, expands the histopathology of resulting lung cancers initiated by Ad-CMV-Cre. The major lung cancer types observed in Ad-SPC-Cre initiated DKO;Ptenflox/flox and DKO;Ptenflox/+ mice were SCLC and AD, with a significant fraction of LCNEC (Figure 2A). Both AD tumor burden and incidence were higher in Ad-SPC-Cre initiated cohorts compared to Ad-CMV-Cre initiated cohorts (Figure 2A), indicating that SPC expressing AT2 cells were more prone to develop AD and less likely to exhibit early development of lethal NE tumors.

Figure 2. Pten loss increases histopathological heterogeneity of lung cancer specimens.

Figure 2.

Open in a new tab

A) The average fractional area of lung tumors exhibiting the indicated histopathological subtypes is shown for mice in the different cohorts. The fractions are rounded to the nearest whole number. B) Images showing representative lung tumor subtypes developing in the mice analyzed, including an image showing 3 subtypes within a single mouse. C) Representative images of lung tissue sections from mice of the four cohorts stained for H&E or immunostained for TTF-1 or SYP. Scale bars represent 2mm. D) Graph depicts the fraction of mice in each cohort developing lung tumors of the indicated subtypes. E) Representative images of AD subtypes observed in the indicated mouse cohorts. F) The table shows the incidence of AD subtypes observed in the indicated mouse cohorts.

We immunostained tissue sections for SYP and TTF-1 to confirm diagnoses (Figure 2C). Nearly 90% of all tumor cells stained positive for TTF-1 in each of the cohorts, distinguishing these tumors from other possible lung cancer types like squamous carcinoma. SYP positive tumors developed in all mice examined while SYP negative AD developed in about two-thirds of mice, particularly in the Ad-SPC-Cre initiated cohorts (Figure 2D). The majority of NE positive tumors in Ad-CMV-Cre initiated mice were located centrally near the major airways, consistent with the anatomical location of most murine pulmonary neuroendocrine cells28. Tumors arising in Ad-SPC-Cre initiated mice were more evenly distributed between central and peripheral locations, consistent with the anatomical location of AT2 cells. We hypothesize that early developing NE tumors located near the major airways explains the shorter lifespan of Ad-CMV-Cre cohorts.

Clinically, major histologic AD subtypes correlate with prognosis46. Lepidic AD is associated with better prognosis in humans, acinar and papillary AD with intermediate prognosis, and solid or micropapillary AD with poor prognosis10, 46–48. Lung cancer GEMMs initiated by Rb1/Trp53 deletion have not been systematically analyzed for analogous AD subtypes, so we addressed this here. AD developing in the four cohorts exhibit a range of different subtypes (Figure 2E). Interestingly, the distribution of AD subtypes varies depending on genotype and the cell of origin (Figure 2F). Nearly all AD developing in Ad-CMV-Cre initiated mice is lepidic, with rarer occurrence of a solid growth pattern (2 of 17 mice). Atypical papillary bronchiolar proliferation (APBP), which has been suggested to be a premalignant lesion36, is also observed in some of these mice. Lepidic and acinar AD are the predominant growth patterns in Ad-SPC-Cre initiated DKO;Ptenflox/flox mice. Ad-SPC-Cre initiated DKO;Ptenflox/+ mice exhibit the broadest range of AD subtypes, including the more aggressive solid and micropapillary subtypes. Acinar, solid, papillary and micropapillary AD is observed in a single mouse surviving 67 weeks post-virus delivery. As the greatest range of AD subtypes are observed in the longest lived AD-SPC-Cre initiated DKO;Ptenflox/+ mice, the detection of these subtypes may be precluded in other cohorts by earlier developing lethal NE tumors.

We have examined tumor tissues collected 6–10 weeks earlier than the median survival time for each cohort to assess cancer histopathological diversity earlier during cancer progression. Tumor burden is similar between cohorts at these earlier times of collection, although there is a wide range of tumor burden in different mice within the same cohort (Figure 3A). We observe clear differences in the distribution of tumor type fractional area between cohorts at the early time points and between the early and late time points for Ad-SPC-Cre initiated mice (Figure 3B). In general, SCLC has reduced incidence (Figure 3C) and makes up a smaller fraction of tumor area at the early time points. In contrast AD makes up a greater fraction of tumor area at these early time points. For example, AD contributes 80% and SCLC 17% to total tumor area at early time points in Ad-SPC-Cre initiated DKO;Ptenflox/flox mice, but 21% and 51% 6–10 weeks later. Tumor sections from Ad-SPC-Cre initiated mice developing both AD and NE tumors were immunostained for the proliferation marker KI67 and the fractional Ki67 positive tumor area compared. Average Ki67 positive tumor area was high for both (Figure 3D), indicative of the high proliferative index expected of tumors cancers lacking Rb1/Trp53. Yet, KI67 immunostaining was significantly higher for NE tumors than for AD (Figure 3E, T test P=0.005). These data indicate that high grade NE tumors likely develop later than AD tumors, but they proliferate and expand faster.

Figure 3. Histopathological heterogeneity of lung cancers at earlier time points.

Figure 3.

Open in a new tab

A) Tumor burden is analyzed 6–10 weeks prior to the median survival age for mice in the indicated cohorts as measured by the fractional area of the lung comprised of neoplastic tissue (not significant by one-way ANOVA P=0.64). B) Lung cancer subtypes prevalent in mice at the earlier time points is shown, as measured in figure 2A. The fractions are rounded to the nearest whole number. C) Graph depicts the fraction of mice at the earlier time points developing lung tumors of the indicated subtypes. D) Lung tissue sections from mice at the early time points were immunostained for KI67 and representative images are shown. Scale bar is 2mm. E) The fraction of KI67 positive tumor area is shown for AD and NE tumors in these mice. Each dot represents an individual mouse. Differences between AD and NE tumors are significant (P<0.01, T test).

Pten loss drives lung cancer molecular heterogeneity.

Synaptophysin (SYP) is a neuroendocrine marker used to diagnose SCLC and other NE tumors. SYP positive tumor foci were detected in a large fraction of mice from all cohorts, although SYP positive fractional area was reduced in the Ad-SPC-Cre initiated cohorts (Figure 4A,B). To investigate this heterogeneity further, we quantitated SYP immunostaining based on intensity (Figure 4C). Most Ad-SPC-Cre initiated tumors exhibited weaker SYP immunostaining, consistent with the greater prevalence of AD compared to Ad-CMV-Cre initiated tumors. Ad-CMV-Cre initiated DKO;Ptenflox/+ mice had the highest fraction of SYP positive tumor area and the highest SYP immunostaining intensity, again matching the tumor subtype prevalence as nearly 90% of tumors arising in this cohort were SCLC. Although we observed heterogeneous SYP expression in primary tumors, all liver metastases developing in mice from the four cohorts exhibited uniformly strong SYP immunostaining and histology consistent with SCLC (Figure 4D). Of the primary tumor subtypes identified, therefore, SCLC has the greatest metastatic potential.

Figure 4. SYP immunostaining is heterogeneous in lung cancer specimens.

Figure 4.

Open in a new tab

A) Lung tissue sections from mice of the indicated cohorts were immunostained for the NE marker SYP, and the mean fraction of SYP positive tumor area is shown (one-way ANOVA P=0.027). B) Representative images of lung tumor sections from the indicated cohorts immunostained for SYP are shown. Scale bar is 200 μm. C) The intensity of SYP immunostaining was assessed on a scale of 0–3, and the fraction of tumor area with the indicated staining intensity is shown in the graph at left, averaged across mice in the indicated cohorts. At right our representative images of SYP immunostained tumor sections indicative of the different immunostaining scores (two-way ANOVA P<0.01). D) Representative images of liver tissue sections from each of the mouse cohorts immunostained for SYP is shown. All mice develop SCLC-like liver metastasis. Scale is 2mm.

RNA-seq analysis has been performed on representative lung tissue from mice in each of the four cohorts to assess transcriptional heterogeneity between mice and cohorts. Principal component analysis and hierarchical clustering indicates that gene expression does not correlate well with genotype or adenovirus treatment, except for the Ad-CMV-Cre initiated DKO;Ptenflox/flox cohort whose samples cluster closely (Figure 5A,B). Gene expression varied considerably among samples within the remaining cohorts, with Ad-SPC-Cre initiated tumors covering a wider area of principal component space (Figure 5C). This is consistent with findings in human lung cancers where gene expression does not correlate well with underlying genetic alterations49. The closely clustering samples from the Ad-CMV-Cre initiated DKO;Ptenflox/flox cohort (3F12, 3F21, 3F30) show reduced NE gene expression and increased immunomodulatory and defense response gene expression relative to other clusters (Figure 5D), functions consistent with normal pulmonary club cell function50. Indeed, these samples uniquely express club cell marker genes (Figure 5E). Another looser cluster of four samples with predominant SCLC pathology (3F10, T1303, T26, T1822) exhibit elevated expression of lineage specifying transcription factors expressed in retinal lineages but not normally in the lung (Lhx1, Dmbx1)(Figure 5F). This group of samples also express higher levels of Mycl and Nfib. Nfib is amplified recurrently in mouse SCLC tumors initiated by Rb1/Trp53 deletion, can drive SCLC development, and has been associated with SCLC metastasis25, 51. Even though two of these samples were initiated with Ad-SPC-Cre, they all exhibit relatively low expression levels of AT2 cell marker genes (Sftpc, Lamp3). SCLC in these four samples is thus molecularly distinct from other samples with predominant SCLC histology (e.g. T1599, T31). All SCLC developing in these mice appear to be of the Ascl1 high subtype given Pou2f3 and NeuroD1 expression is low or undetectable (Table S1). Yap1 expression is highest in the club cell-like cluster of samples. Thus there is considerable transcriptional heterogeneity within the Ascl1 high NE tumor subtype.

Figure 5. Gene expression heterogeneity in lung cancer specimens.

Figure 5.

Open in a new tab

A) End stage lung tissue from 3 mice in each cohort was analyzed by RNA-seq. The graph shows principal component analysis with individual samples shown, coded for genotype and adenovirus treatment. B) Samples were analyzed by IQR hierarchical clustering. The graph shows clustering by the top 300 most informative genes. Sample names are color coded by clusters noted in text (red=club cell like, blue=Nfib+). C) The fractional area of lung tumors exhibiting the indicated histopathological subtypes is shown for each mouse used in RNA-seq analysis. D) GSEA analysis of the cluster comprised of samples 3F12, 3F21, and 3F30 showing up and down regulated gene sets with the adjusted P value. E) Graph depicts relative RNA-seq read counts for the indicated club cell lineage marker genes. Each dot represents an individual sample, color coded as in C. F) Graph depicts relative RNA-seq counts for the indicated genes, as in D.

End stage lung tissues analyzed above are likely composed of multiple tumor foci, thus the bulk RNA-seq data reflects potential admixtures of tumor subtype heterogeneity prevalent in individual tissues analyzed. To account for this, we have performed spatial transcriptomic profiling of tissue sections from the same mice. The 12 samples analyzed in aggregate generated 39 distinguishable transcriptional clusters (Figure 6A). Consistent with bulk RNA-seq analysis, most clusters are private to a particular sample indicating substantial gene expression heterogeneity between mice even within a cohort (Figure 6B). The exception is the Ad-CMV-Cre initiated DKO:Ptenflox/flox cohort whose samples cluster more closely together. An NE signature score was calculated for each spot on the spatial transcriptomic profiles, and this score was mapped to the gene expression clusters (Figure 6C). The NE score varied considerably between and within samples; the predominant adenocarcinoma sample (T1821) shows the lowest NE score, as expected. The club cell-like clusters exhibit relatively low NE scores consistent with bulk RNA-seq data. Samples with predominant SCLC histology vary in NE score considerably, with the four Nfib+ samples (3F10, T1303, T1822, T26) expressing lower NE scores than other samples with predominant SCLC histology (T1303, T1599). Because of the transcriptional heterogeneity observed, more experiments and analyses will be required to test if histopathologic subtypes share a common gene expression program.

Figure 6. Spatial transcriptomic profiling of lung cancer specimens.

Figure 6.

Open in a new tab

A) Spatial transcriptomic profiling data from twelve lung cancer tissue samples analyzed in figure 5 was analyzed in aggregate by Uniform Manifold Approximation and Projection. Clusters distinguishable by RNA expression patterns are color coded and numbered. B) Analysis as in A) but color coded by sample identity. The sample font color is coded by genotype and cell of origin (CMV:Ptenflox/flox=red, CMV:Ptenflox/+=green, SPC:Ptenflox/flox=blue, SPC:Ptenflox/+=black). The cohort whose samples cluster together is outlined. C) A neuroendocrine signature score was calculated for each spot on the spatial profiles of all twelve samples and the score mapped to the UMAP clusters in A). D) Spatial transcriptomic data from sample T1297 alone was analyzed as in A) to generate higher resolution clustering. E) The image at left is the H&E stained tissue section from sample T1297 used for spatial transcriptomic profiling. The middle image shows the transcriptional clusters spatially mapped to the section. The image at right maps the NE gene expression score from C) to the tissue section.

Mapping the 39 transcriptional clusters to the tissue sections indicates these transcriptional clusters are typically spatially distinct with some limited spatial intermixing (Figure S1). The Ad-CMV-Cre initiated DKO:Ptenflox/flox cohort contains the tumor cluster with the most significantly elevated club cell marker gene expression (cluster 14- Table S2). This cluster is shared across the different samples in this cohort. Further the histopathology corresponding to this cluster is predominantly LCNEC across all samples in this cohort, demonstrating reproducibility. Including only tumor area, Ad-SPC-Cre initiated DKO:Ptenflox/flox samples expressed more distinct transcriptional clusters (9) than Ad-SPC-Cre initiated DKO:Ptenflox/+ samples (5).

When samples are analyzed individually, more transcriptional clusters are detectable per sample. For example, fifteen distinct transcriptional clusters are distinguished in sample T1297 that exhibits predominant LCNEC histology (Figure 6D, S2). Mapping these more highly resolved transcriptional clusters spatially also indicates limited intermixing (Figure 6E). Tumor foci alone encompass at least 7 of the 15 transcriptional clusters identified. The NE gene expression score varies both between and within spatially separated tumor foci. There is evidence that cells near the periphery of tumor foci have distinct gene expression profiles compared to cells near the center of tumor foci (cluster 0), perhaps reflecting effects of distinct cell microenvironments. Some transcriptional clusters are unique to an area within an individual tumor foci (clusters 11, 14). These two clusters retain gene expression characteristic of AT2 cells as they are initiated by Ad-SPC-Cre, they have relatively low NE gene expression score, and they significantly differentially express genes (adj. P<0.05) normally expressed in cell lineages outside of lung epithelium. Examples of alternative lineage gene expression in these clusters include mesenchymal (Wnt11, Serpine1, Cldn11), squamous (Calml3, Ivl), retinal (Vgf), glandular (Upk3a), and intestinal (Nkx2-2)(Table S3). Overall the gene expression analysis suggests complete Pten loss expands the lineage plasticity of resulting lung tumors.

Discussion

Detailed histopathological characterization and spatial transcriptomic profiling of new mouse models has been conducted to test how Pten loss influences the phenotype of lung cancers initiated by Rb1/Trp53 deletion. PTEN loss of function is a clinically relevant genetic alteration as it is recurrent in human SCLC (~6% of cases, cBioPortal). Further, Pten loss of function alterations arise spontaneously and recurrently in mouse models of SCLC initiated by Rb1/Trp53 deletion25, suggesting Pten loss of function is selected during tumor evolution. We confirm this hypothesis here as complete Pten deletion accelerates lung tumorigenesis initiated by Rb1/Trp53 loss thus shortening mouse survival compared to mice lacking only one Pten allele. Results are consistent with prior published work where Rb1/Trp53/Pten were deleted in CGRP expressing pulmonary neuroendocrine cells or using Ad-CMV-Cre25, 26. Our results extend this conclusion to SPC expressing AT2 cells, the presumed cell of origin for most lung adenocarcinomas. The median survival of Ad-SPC-Cre initiated DKO;Ptenflox/flox (259 days) and DKO;Ptenflox/+ (364 days) mice is significantly shorter than that reported previously for Ad-SPC-Cre initiated mice lacking Rb1/Trp53 alone (462 days)28. Overall lung cancer penetrance is also higher (100% vs. 73%). While not tested directly here, these findings suggest loss of one Pten allele may also accelerate lung tumorigenesis and shorten mouse survival.

Earlier work did not characterize the histopathology of lung tumors developing upon Pten/Rb1/Trp53 deletion nor were these alterations restricted to AT2 cells. Through detailed histopathological characterization of lung tumors initiated with these genetic alterations in AT2 cells (Ad-SPC-Cre), we can conclude that deleting Pten/Rb1/Trp53 expands the phenotypic heterogeneity of resulting lung cancers. For example, Rb1/Trp53 deletion in AT2 cells has been reported to generate SCLC with histopathological and molecular features analogous to SCLC arising from pulmonary neuroendocrine cells28. In contrast, we observe that Pten/Rb1/Trp53 deletion in AT2 cells generates a substantial fraction of LCNEC, AD, and AD-NE. Some mice develop LCNEC predominantly, making these mice particularly useful for studying this disease state given the paucity of available LCNEC mouse models. Histopathological diversity within individual mice is greater upon complete loss of Pten/Rb1/Trp53 in AT2 cells compared to AT2 cells retaining one wild type Pten allele. A markedly higher incidence of AD is observed in AD-SPC-Cre initiated cohorts studied here compared to prior published studies deleting Rb1/Trp53 alone (80% vs. 3%), suggesting Pten deficiency increased the likelihood initiated AT2 cells develop AD. The diversity of AD pathological subtypes arising varies depending on cell of origin and Pten allele dosage. The most diverse AD histopathology is observed in the longest-lived mice, likely because AD is slower growing. There is substantial mouse to mouse variation both between and within mouse cohorts with respect to tumor burden and histopathology. Overall these observations are consistent with the hypothesis that Pten loss increases the susceptibility of Rb1/Trp53 deficient AT2 cells to neoplastic transformation and drives the phenotypic heterogeneity of resulting tumors.

Spatial transcriptomic profiling demonstrates significant molecular heterogeneity in tumors developing in mice between different cohorts, between mice within the same cohort, between spatially distinct tumor foci within individual mice, and within individual tumor foci. Lineage plasticity is elevated in these tumors with expression of genes normally restricted to lineages outside of the lung. Given this transcriptional plasticity, the tumor cell growth environment likely influences phenotype. Spatial profiling provides evidence in support of this hypothesis as the transcriptional patterns cluster based on the tumor cell’s position with limited spatial intermixing. For example, cells at the periphery of a tumor foci can have a transcriptional pattern distinct from cells at the center of the foci. NE tumors arising in Ad-CMV-Cre initiated DKO;Ptenflox/flox mice are an exception since their gene expression is similar between mice. These tumors uniquely express genes normally restricted to the club cell lineage. While lineage tracing approaches are required to rigorously test this, the finding suggests club cells are susceptible to infection by Ad-CMV-Cre and can serve as the cell of origin for these lung tumors. Tumors form rarely when Rb1/Trp53 alone are deleted in club cells specifically28, implying that complete loss of Pten can increase the susceptibility of club cells to neoplastic transformation. Alternatively, Pten loss may drive lung cancer lineage plasticity sufficiently to allow reprogramming to a club cell like transcriptional phenotype, although this is not observed when tumors are initiated with Ad-SPC-Cre.

Genetically engineered mouse models have been instrumental in deciphering the contributions of genetic alterations and the cell of origin to the development of lung cancers in general52 and SCLC in particular23, 29. Results here extend this body of literature by demonstrating that Pten deficiency, when combined with Rb1/Trp53 alteration, can drive both more efficient neoplastic transformation of AT2 cells and elevated cancer lineage plasticity. We note that similar genetic interactions between Rb1, Trp53, and Pten in driving neoplastic transformation and lineage plasticity has also been observed in prostate cancer34, 53. The diversity of tumor phenotypes arising in these new mouse models will be potentially useful for future studies aimed at understanding how genetic mutations and the cell or origin collaborate to influence lung cancer heterogeneity and assessing the relative sensitivity of different lung cancer subtypes to cancer therapy.

Supplementary Material

1

Figure S1. Spatial heterogeneity in lung cancer gene expression. Transcriptional clusters identified in figure 6 were mapped to the tissue sections from which the data was generated. For each sample, the image on left is the H&E-stained tissue section while the image on the right shows the transcriptional cluster mapping. Transcriptional clusters are typically private to individual samples and spatially distinct with limited spatial intermixing.

2

Figure S2. Genes contributing to transcriptional clustering for sample T1297. A heatmap of gene expression markers that distinguish between transcriptional clusters identified from spatial profiling data described in figure 6D,E.

3

Table S1. Normalized RNA-seq read counts from gene expression analysis of bulk tumors. The table lists the normalized gene-read count matrix for the RNA-seq data generated from 3 samples for each cohort. TKO indicates the Ptenflox/flox:Rb1flox/flox:Trp53flox/flox genotype while DKO is the Ptenflox/+:Rb1flox/flox:Trp53flox/flox genotype. CMV indicates initiation with Ad-CMV-Cre, while SPC indicates initiation with Ad-SPC-Cre. Individual sample numbers are listed corresponding to the figures.

4

Table S2. Spatial transcriptomic clusters identified in aggregate from all samples. The table lists the 39 transcriptional clusters detected when analyzing spatial transcriptomic data from all samples in aggregate. Genes significantly differentially expressed are listed for each cluster with P value, adjusted P value, average log2 fold change, and percent cells where expression is detected in the cluster (pct.1) and outside the cluster (pct.2).

5

Table S3. Spatial transcriptomic clusters identified within individual samples. For each sample the table lists the transcriptional clusters detected within that sample. Genes significantly differentially expressed are listed for each cluster with P value, adjusted P value, average log2 fold change, and percent cells where expression is detected in the cluster (pct.1) and outside the cluster (pct.2).

Acknowledgements

This research was supported by grants to D.W.G. (NCI R01 CA234162, NCI R01 CA207757, and the Roswell Park Alliance Foundation) and X.Z. (National Natural Science Foundation of China: 81300045). This work was supported by the Roswell Park Comprehensive Cancer Center and the National Cancer Institute (NCI) grant P30 CA016056.

Conflict of Interest Disclosure:

The authors do not declare any conflict of interest. Support for the research was provided by the National Institutes of Health (USA), the Roswell Park Alliance Foundation, and the National Natural Science Foundation of China.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  • 1.Gazdar AF, Bunn PA, Minna JD. Small-cell lung cancer: what we know, what we need to know and the path forward. Nat Rev Cancer 2017;17:765. [DOI] [PubMed] [Google Scholar]
  • 2.Lally BE, Urbanic JJ, Blackstock AW, et al. Small cell lung cancer: have we made any progress over the last 25 years? Oncologist 2007;12:1096–1104. [DOI] [PubMed] [Google Scholar]
  • 3.Rudin CM, Brambilla E, Faivre-Finn C, et al. Small-cell lung cancer. Nat Rev Dis Primers 2021;7:3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Bottger F, Semenova EA, Song JY, et al. Tumor Heterogeneity Underlies Differential Cisplatin Sensitivity in Mouse Models of Small-Cell Lung Cancer. Cell reports 2019;27:3345–3358 e3344. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Shue YT, Lim JS, Sage J. Tumor heterogeneity in small cell lung cancer defined and investigated in pre-clinical mouse models. Transl Lung Cancer Res 2018;7:21–31. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Travis WD. Update on small cell carcinoma and its differentiation from squamous cell carcinoma and other non-small cell carcinomas. Mod Pathol 2012;25 Suppl 1:S18–30. [DOI] [PubMed] [Google Scholar]
  • 7.van Meerbeeck JP, Fennell DA, De Ruysscher DK. Small-cell lung cancer. Lancet 2011;378:1741–1755. [DOI] [PubMed] [Google Scholar]
  • 8.Nicholson SA, Beasley MB, Brambilla E, et al. Small cell lung carcinoma (SCLC): a clinicopathologic study of 100 cases with surgical specimens. Am J Surg Pathol 2002;26:1184–1197. [DOI] [PubMed] [Google Scholar]
  • 9.Raso MG, Bota-Rabassedas N, Wistuba II. Pathology and Classification of SCLC. Cancers (Basel) 2021;13. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Travis WD, Brambilla E, Nicholson AG, et al. The 2015 World Health Organization Classification of Lung Tumors: Impact of Genetic, Clinical and Radiologic Advances Since the 2004 Classification. Journal of thoracic oncology : official publication of the International Association for the Study of Lung Cancer 2015;10:1243–1260. [DOI] [PubMed] [Google Scholar]
  • 11.Niederst MJ, Sequist LV, Poirier JT, et al. RB loss in resistant EGFR mutant lung adenocarcinomas that transform to small-cell lung cancer. Nature communications 2015;6:6377. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Sequist LV, Waltman BA, Dias-Santagata D, et al. Genotypic and histological evolution of lung cancers acquiring resistance to EGFR inhibitors. Science translational medicine 2011;3:75ra26. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Offin M, Chan JM, Tenet 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. Journal of thoracic oncology : official publication of the International Association for the Study of Lung Cancer 2019;14:1784–1793. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Oser MG, Niederst MJ, Sequist LV, et al. Transformation from non-small-cell lung cancer to small-cell lung cancer: molecular drivers and cells of origin. Lancet Oncol 2015;16:e165–172. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Rudin CM, Poirier JT, Byers LA, et al. Molecular subtypes of small cell lung cancer: a synthesis of human and mouse model data. Nat Rev Cancer 2019. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Schwendenwein A, Megyesfalvi Z, Barany N, et al. Molecular profiles of small cell lung cancer subtypes: therapeutic implications. Mol Ther Oncolytics 2021;20:470–483. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Poirier JT, George J, Owonikoko TK, et al. New Approaches to SCLC Therapy: From the Laboratory to the Clinic. Journal of thoracic oncology : official publication of the International Association for the Study of Lung Cancer 2020;15:520–540. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.George J, Lim JS, Janq SJ, et al. Comprehensive genomic profiles of small cell lung cancer. Nature 2015. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Peifer M, Fernandez-Cuesta L, Sos ML, et al. Integrative genome analyses identify key somatic driver mutations of small-cell lung cancer. Nat Genet 2012;44:1104–1110. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Rudin CM, Durinck S, Stawiski EW, et al. Comprehensive genomic analysis identifies SOX2 as a frequently amplified gene in small-cell lung cancer. Nat Genet 2012;44:1111–1116. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Gardner EE, Lok BH, Schneeberger VE, et al. Chemosensitive Relapse in Small Cell Lung Cancer Proceeds through an EZH2-SLFN11 Axis. Cancer cell 2017;31:286–299. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Gazdar AF, Savage TK, Johnson JE, et al. The comparative pathology of genetically engineered mouse models for neuroendocrine carcinomas of the lung. Journal of thoracic oncology : official publication of the International Association for the Study of Lung Cancer 2015;10:553–564. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Drapkin BJ, Rudin CM. Advances in Small-Cell Lung Cancer (SCLC) Translational Research. Cold Spring Harb Perspect Med 2021;11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Meuwissen R, Linn SC, Linnoila RI, et al. Induction of small cell lung cancer by somatic inactivation of both Trp53 and Rb1 in a conditional mouse model. Cancer cell 2003;4:181–189. [DOI] [PubMed] [Google Scholar]
  • 25.McFadden DG, Papagiannakopoulos T, Taylor-Weiner A, et al. Genetic and clonal dissection of murine small cell lung carcinoma progression by genome sequencing. Cell 2014;156:1298–1311. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Cui M, Augert A, Rongione M, et al. PTEN is a potent suppressor of small cell lung cancer. Mol Cancer Res 2014;12:654–659. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Ouadah Y, Rojas ER, Riordan DP, et al. Rare Pulmonary Neuroendocrine Cells Are Stem Cells Regulated by Rb, p53, and Notch. Cell 2019;179:403–416 e423. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Sutherland KD, Proost N, Brouns I, et al. Cell of origin of small cell lung cancer: inactivation of Trp53 and rb1 in distinct cell types of adult mouse lung. Cancer cell 2011;19:754–764. [DOI] [PubMed] [Google Scholar]
  • 29.Ferone G, Lee MC, Sage J, et al. Cells of origin of lung cancers: lessons from mouse studies. Genes Dev 2020;34:1017–1032. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Park KS, Liang MC, Raiser DM, et al. Characterization of the cell of origin for small cell lung cancer. Cell Cycle 2011;10:2806–2815. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Yang D, Denny SK, Greenside PG, et al. Intertumoral Heterogeneity in SCLC Is Influenced by the Cell Type of Origin. Cancer discovery 2018;8:1316–1331. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Wang S, Gao J, Lei Q, et al. Prostate-specific deletion of the murine Pten tumor suppressor gene leads to metastatic prostate cancer. Cancer cell 2003;4:209–221. [DOI] [PubMed] [Google Scholar]
  • 33.Sun H, Wang Y, Chinnam M, et al. E2f binding-deficient Rb1 protein suppresses prostate tumor progression in vivo. Proc Natl Acad Sci U S A 2011;108:704–709. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Ku SY, Rosario S, Wang Y, et al. Rb1 and Trp53 cooperate to suppress prostate cancer lineage plasticity, metastasis, and antiandrogen resistance. Science 2017;355:78–83. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.DuPage M, Dooley AL, Jacks T. Conditional mouse lung cancer models using adenoviral or lentiviral delivery of Cre recombinase. Nat Protoc 2009;4:1064–1072. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Iwanaga K, Yang Y, Raso MG, et al. Pten inactivation accelerates oncogenic K-ras-initiated tumorigenesis in a mouse model of lung cancer. Cancer Res 2008;68:1119–1127. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Andrews S FASTQC. A quality control tool for high throughput sequence data. 2010. Available at https://www.bioinformatics.babraham.ac.uk/projects/fastqc/.
  • 38.Kent WJ, Sugnet CW, Furey TS, et al. The human genome browser at UCSC. Genome Res 2002;12:996–1006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Langmead B, Salzberg SL. Fast gapped-read alignment with Bowtie 2. Nature methods 2012;9:357–359. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Kim D, Pertea G, Trapnell C, et al. TopHat2: accurate alignment of transcriptomes in the presence of insertions, deletions and gene fusions. Genome biology 2013;14:R36. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Wang L, Wang S, Li W. RSeQC: quality control of RNA-seq experiments. Bioinformatics 2012;28:2184–2185. [DOI] [PubMed] [Google Scholar]
  • 42.Liao Y, Smyth GK, Shi W. featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics 2014;30:923–930. [DOI] [PubMed] [Google Scholar]
  • 43.Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome biology 2014;15:550. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Kolde R pheatmap: Pretty Heatmaps. 2015. Available at https://rdrr.io/cran/pheatmap/.
  • 45.Liberzon A, Subramanian A, Pinchback R, et al. Molecular signatures database (MSigDB) 3.0. Bioinformatics 2011;27:1739–1740. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Zheng M. Classification and Pathology of Lung Cancer. Surg Oncol Clin N Am 2016;25:447–468. [DOI] [PubMed] [Google Scholar]
  • 47.Kadota K, Kushida Y, Kagawa S, et al. Cribriform Subtype is an Independent Predictor of Recurrence and Survival After Adjustment for the Eighth Edition of TNM Staging System in Patients With Resected Lung Adenocarcinoma. Journal of thoracic oncology : official publication of the International Association for the Study of Lung Cancer 2019;14:245–254. [DOI] [PubMed] [Google Scholar]
  • 48.Tsutsumida H, Nomoto M, Goto M, et al. A micropapillary pattern is predictive of a poor prognosis in lung adenocarcinoma, and reduced surfactant apoprotein A expression in the micropapillary pattern is an excellent indicator of a poor prognosis. Mod Pathol 2007;20:638–647. [DOI] [PubMed] [Google Scholar]
  • 49.Tang M, Abbas HA, Negrao MV, et al. The histologic phenotype of lung cancers is associated with transcriptomic features rather than genomic characteristics. Nature communications 2021;12:7081. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Dean CH, Snelgrove RJ. New Rules for Club Development: New Insights into Human Small Airway Epithelial Club Cell Ontogeny and Function. Am J Respir Crit Care Med 2018;198:1355–1356. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 51.Semenova EA, Kwon MC, Monkhorst K, et al. Transcription Factor NFIB Is a Driver of Small Cell Lung Cancer Progression in Mice and Marks Metastatic Disease in Patients. Cell reports 2016;16:631–643. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52.Rowbotham SP, Kim CF. Diverse cells at the origin of lung adenocarcinoma. Proc Natl Acad Sci U S A 2014;111:4745–4746. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53.Mu P, Zhang Z, Benelli M, et al. SOX2 promotes lineage plasticity and antiandrogen resistance in TP53- and RB1-deficient prostate cancer. Science 2017;355:84–88. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1

Figure S1. Spatial heterogeneity in lung cancer gene expression. Transcriptional clusters identified in figure 6 were mapped to the tissue sections from which the data was generated. For each sample, the image on left is the H&E-stained tissue section while the image on the right shows the transcriptional cluster mapping. Transcriptional clusters are typically private to individual samples and spatially distinct with limited spatial intermixing.

NIHMS1861659-supplement-1.pdf (8.3MB, pdf)
2

Figure S2. Genes contributing to transcriptional clustering for sample T1297. A heatmap of gene expression markers that distinguish between transcriptional clusters identified from spatial profiling data described in figure 6D,E.

NIHMS1861659-supplement-2.pdf (339.4KB, pdf)
3

Table S1. Normalized RNA-seq read counts from gene expression analysis of bulk tumors. The table lists the normalized gene-read count matrix for the RNA-seq data generated from 3 samples for each cohort. TKO indicates the Ptenflox/flox:Rb1flox/flox:Trp53flox/flox genotype while DKO is the Ptenflox/+:Rb1flox/flox:Trp53flox/flox genotype. CMV indicates initiation with Ad-CMV-Cre, while SPC indicates initiation with Ad-SPC-Cre. Individual sample numbers are listed corresponding to the figures.

NIHMS1861659-supplement-3.xlsx (3.8MB, xlsx)
4

Table S2. Spatial transcriptomic clusters identified in aggregate from all samples. The table lists the 39 transcriptional clusters detected when analyzing spatial transcriptomic data from all samples in aggregate. Genes significantly differentially expressed are listed for each cluster with P value, adjusted P value, average log2 fold change, and percent cells where expression is detected in the cluster (pct.1) and outside the cluster (pct.2).

NIHMS1861659-supplement-4.xlsx (9.3MB, xlsx)
5

Table S3. Spatial transcriptomic clusters identified within individual samples. For each sample the table lists the transcriptional clusters detected within that sample. Genes significantly differentially expressed are listed for each cluster with P value, adjusted P value, average log2 fold change, and percent cells where expression is detected in the cluster (pct.1) and outside the cluster (pct.2).

NIHMS1861659-supplement-5.xlsx (21.3MB, xlsx)

RESOURCES