(C) PLOS One [1]. This unaltered content originally appeared in journals.plosone.org.
Licensed under Creative Commons Attribution (CC BY) license.
url:https://journals.plos.org/plosone/s/licenses-and-copyright

------------



Epithelial-specific ERBB3 deletion results in a genetic background-dependent increase in intestinal and colon polyps that is mediated by EGFR

['Carolina Mantilla Rojas', 'Interdisciplinary Program In Genetics', 'Texas A M University', 'College Station', 'Texas', 'United States Of America', 'Department Of Molecular', 'Cellular Medicine', 'Michael P. Mcgill', 'Anna C. Salvador']

Date: 2022-02 

Abstract ERBB3 has gained attention as a potential therapeutic target to treat colorectal and other types of cancers. To confirm a previous study showing intestinal polyps are dependent upon ERBB3, we generated an intestinal epithelia-specific ERBB3 deletion in C57BL/6-ApcMin/+ mice. Contrary to the previous report showing a significant reduction in intestinal polyps with ablation of ERBB3 on a B6;129 mixed genetic background, we observed a significant increase in polyp number with ablation of ERBB3 on C57BL/6J compared to control littermates. We confirmed the genetic background dependency of ERBB3 by also analyzing polyp development on B6129 hybrid and B6;129 advanced intercross mixed genetic backgrounds, which showed that ERBB3 deficiency only reduced polyp number on the mixed background as previously reported. Increased polyp number with ablation of ERBB3 was also observed in C57BL/6J mice treated with azoxymethane showing the effect is model independent. Polyps forming in absence of ERBB3 were generally smaller than those forming in control mice, albeit the effect was greatest in genetic backgrounds with reduced polyp numbers. The mechanism for differential polyp number in the absence of ERBB3 was through altered proliferation. Backgrounds with increased polyp number with loss of ERBB3 showed an increase in cell proliferation even in non-tumor epithelia, while backgrounds showing reduced polyp number with loss of ERBB3 showed reduced cellular proliferation. Increase polyp number caused by loss of ERBB3 was mediated by increased epidermal growth factor receptor (EGFR) expression, which was confirmed by deletion of Egfr. Taken together, this study raises substantial implications on the use of ERBB3 inhibitors against colorectal cancer. The prediction is that some patients may have increased progression with ERBB3 inhibitor therapy, which is consistent with observations reported for ERBB3 inhibitor clinical trials.

Author summary Targeted cancer therapeutics that are efficacious in pre-clinical studies do not always translate to the clinic. This can be due to the homogenous pre-clinical models not representing heterogeneous patient populations. Small molecular inhibitors to ERBB3 for colorectal cancer efficiently inhibited tumorigenesis in pre-clinical studies, but in clinical trials failed to show efficacy and even trended toward enhancing cancer growth. Using two mouse colorectal cancer models with variable the genetic backgrounds, we show that epithelial deletion of ERBB3 can result in outcomes ranging from inhibition to enhanced tumor growth. Molecular profiling implicated EGFR as a likely mediator of enhanced tumor growth in the absence of ERBB3. Mice with epithelial deletion of both EGFR and ERBB3 were used to show that the background dependency of ERBB3 is mediated by EGFR. These results show that genetic context of cancer therapeutics can have profound implications on translating pre-clinical results into the clinic.

Citation: Mantilla Rojas C, McGill MP, Salvador AC, Bautz D, Threadgill DW (2021) Epithelial-specific ERBB3 deletion results in a genetic background-dependent increase in intestinal and colon polyps that is mediated by EGFR. PLoS Genet 17(11): e1009931. https://doi.org/10.1371/journal.pgen.1009931 Editor: Kent Hunter, National Cancer Institute, UNITED STATES Received: May 21, 2021; Accepted: November 5, 2021; Published: November 29, 2021 Copyright: © 2021 Mantilla Rojas et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability: All relevant data are within the manuscript and its Supporting Information files. Funding: This work was supported by National Institutes of Health Grants R01 CA092479 and NIEHS P30 ES029067, and the Tom and Jean McMullin Chair of Genetics to D.W.T. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing interests: The authors have declared that no competing interests exist.

Introduction Members of the EGFR/ERBB receptor tyrosine family have been studied intensively as therapeutic targets against a variety of cancers [1]. Because of its unique characteristics, attention has focused on ERBB3, which shares core structural domains with other ERBB receptors [2]. However, ERBB3 cannot auto-phosphorylate due to a lack of intrinsic kinase activity distinguishing it from other ERBB receptors [3], although recent evidence suggests the ERBB3 kinase may still be important during heterodimerization with other ERBB receptors [4]. Tyrosine-phosphorylated ERBB3 provides an efficient docking site for downstream adaptor proteins. Notably, ERBB3 has nine docking sites for the p85 subunit of phosphatidylinositol 3-kinase (PI3K) giving it the highest binding affinity for PI3K among the ERBB receptors [5]. Because of the high PI3K binding capacity, protein kinase B (AKT) signaling is strongly activated by ERBB3, which results in an oncogenic stimulus frequently implicated in many cancers and during therapeutic resistance [6]. Although most clinical focus has been on EGFR and ERBB2 due to their aberrant activation in many human malignancies [7], overexpression of ERBB3 often co-occurs with EGFR or ERBB2 in many types of cancers such as breast [8,9], colorectal (CRC) [2,10], gastric [11,12], ovarian [13], and pancreatic [14]. In ERBB2-driven cancers, ERBB3 enhances cancer progression, usually by PI3K/AKT pathway activation, by functioning as a heterodimer partner [15,16]. For these cancers in particular, ERBB3 inhibition may be required to effectively eradicate cancerous cells [17]. Similar to ERBB2, EGFR can be efficiently coupled to the PI3K/AKT pathway by heterodimerization with ERBB3. This is most recognized in non-small cell lung cancers (NSCLC) sensitive to EGFR inhibitors [18]. ERBB3-dependent activation of PI3K/AKT can underlie acquired resistance to EGFR inhibitors in a subset of NSCLC patients that is due to amplification of MET [18]. Blocking of ERBB3 genetically or pharmacologically showed promising preclinical results [19–23], including as a prophylactic vaccination target for at-risk CRC populations [24]. However, in the clinic, surprisingly little evidence for efficacy was observed against CRC [25,26]. A major weakness in most preclinical studies is the lack of attention to variation in genetic background and how this might impact perceived efficacy. This is particularly of concern for clinical trials that are designed primarily to observe off-target toxicities in phase I and progression-free survival in phase II/II. Results of clinical trials with ERBB3 inhibitors against CRC have been disappointing, and possibly even detrimental. The FOCUS4-D trial with AZD8931, a pan-ERBB inhibitor against EGFR, ERBB2, and ERBB3, showed no efficacy, and even trended to shorter progression-free survival over the placebo group [27]. Similarly, comparison of duligotuzumab, an EGFR/ERBB3 dual inhibitor, showed no benefit over the EGFR inhibitor cetuximab and also trended to a lower objective response rate [28]. These results raise concerns whether inhibiting ERBB3 during CRC treatment may actually be detrimental in some patients, masking those cases that might benefit in clinical trials. We previously generated a conditional Erbb3 allele (Erbb3tm1.1Dwt referred as Erbb3f) to perform tissue-specific ablation of ERBB3, which was used to show ERBB3-dependent signaling during intestinal tumorigenesis [23]. On a mixed C57BL/6J and 129S1/SvImJ genetic background (B6;129), we showed that ERBB3 is essential for tumor development. In light of the conflicting clinical trials with ERBB3 inhibition, we attempted to confirm the previous results using a congenic C57BL/6J genetic background. Surprisingly, in two mouse models of human CRC, the ApcMin model of spontaneous intestinal tumorigenesis and a model of colitis-associated colorectal (CAC) tumorigenesis induced by the colon-selective carcinogen azoxymethane (AOM), ERBB3 deficiency dramatically increased tumor multiplicity. Polyps arising in the absence of ERBB3 also showed increased transcript levels of Erbb4 and Egf, consistent with previous results on the B6;129 mixed background [23]. To confirm if the original results were due to the 129S1/SvImJ background, we crossed the conditional Erbb3 allele to the 129S1/SvImJ background and found that ERBB3-deficiency results in a dramatic reduction in polyp multiplicity in the ApcMin/+ mouse model, similar to the original observations. In contrast, on a B6129 hybrid background, the lack of ERBB3 has no impact on tumorigenesis. Based on the known background sensitivity of EGFR, deletion of both Egfr and Erbb3 on a C57BL/6J background showed a dramatic decrease in polyp number indicating the tumor increase with loss of ERBB3 on the C57BL/6J background was at least partially mediated by EGFR. Together these results suggest that ERBB3 inhibitor response in CRC patients could range from highly efficacious to promoting progression depending on the genetic background of the patient, requiring a re-evaluation of how clinical trials are monitored and analyzed.

Discussion ERBB3 is the only member of the ERBB family that lacks intrinsic kinase activity, although evidence shows that tumor phenotypes can be modulated by activation of ERBB3-dependent pathways [38,39]. We examined intestinal polyp development in the absence of ERBB3 activity using intestinal epithelial-specific deletion of Erbb3. Previous analysis using the ApcMin/+ mouse model reported that ERBB3 deficiency significantly decreases mean polyp number and size on a B6;129 mixed genetic background. When we attempted to confirm these results on an isogenic C57BL/6J background, absence of ERBB3 actually has the opposite effect, with a significant increase in the number of polyps. ERBB3-deficiency on a B6129 hybrid background had no effect on polyp number, while a B6;129 mixed background confirmed the previous mixed background results with a significant reduction in the number of polyps. This result demonstrates a striking strain-dependent effect of ERBB3 on the development of intestinal polyps. Although background strongly influences polyp number in the absence of ERBB3, previous prophylactic vaccination studies using a specific anti-ERBB epitope in C57BL/6J-ApcMin/+, which resulted in reduced polyp number [24], indicates that the effect is more complicated and different methods of inhibiting ERBBs may have varying responses. Due to this background dependency, which is reminiscent of the background sensitivity of EGFR on polyp induction [37], we further show that the enhanced polyp number on the C57BL/6J background can be eliminated in a double deficiency of EGFR and ERBB3 in the intestinal epithelium. Other backgrounds may have different compensatory mechanisms when ERBB3 is inhibited based on the heterogenous compensatory mechanisms underlying loss of EGFR [35]. The role of ERBB3 in colonic tumor development was confirmed using a distinctly different model, the AOM mouse model of CRC. These results raise an intriguing dilemma for therapeutic development, the possibility that therapeutic interventions designed to reduce growth of cancers may in fact enhance growth in some patients. Patients that progress under therapy are usually assumed to be non-responders. However, some cancers may have accelerated growth caused by the therapy, which has important implications for inhibitor therapies in the clinic. Despite the genetic background dependency on polyp development, ERBB3-deficiency generally decreased the average size of ApcMin/+ polyps independent of the genetic background, albeit the effect was inconsistent across intestinal regions on backgrounds where ERBB3 deficiency did not reduce polyp number. The effect of reduced in polyp size with ERBB3 loss is opposite to the effect previously reported for loss of EGFR. Previously, Roberts et al., demonstrated that reduced EGFR activity promoted the development of larger intestinal tumors [40], and using the same conditional knockout allele of EGFR used here, we also confirmed that the EGFR-independent intestinal and colonic tumors are larger [41], highlighting the unique role of ERBB3-dependent signaling in regulating tumor growth. To elucidate the mechanism underlying the differentially effect of ERBB3 loss on polyp number and size, we investigated cell proliferation and death in polyps and adjacent normal epithelia. Similar to the previous study on a B6;129 mixed genetic background, a higher number of apoptotic cells were detected in ERBB3-deficient polyps from C57BL/6J-ApcMin/+ mice demonstrating an important role for ERBB3 in tumor cell survival. Transcriptomic analysis also predicted that ERBB3-deficient polyps have decreased levels of p42/44 MAPK activation, which is the predominant mitogenic signal [42]. However, immunohistochemistry results indicate that lack of ERBB3 signaling primarily contributes to tumor growth by increasing cell proliferation as shown in an increase in Ki67 positive cells in ERBB3-deficient polyps. Based on previous studies in lung cancer, PI3K signaling could lead to increase cell proliferation even in the absence of ERBB3 [20,43]. Therefore, the requirement for the ERBB3 signaling pathway in intestinal tumor progression could result from its unique role of linking EGFR signaling to PI3K/AKT, thus activating the PI3K pathway to promote cell proliferation. EGFR can also activate PI3K/AKT through association with the adaptor protein GAB1 in ApcMin/+ polyps [44]. It is possible that PI3K/AKT is activated by EGFR via two mechanisms, association with GAB1 and coupling with ERBB3. Based on previous data and results here, we propose that ERBB3 activation of PI3K/AKT is one of the primary mechanisms supporting ApcMin/+ polyps. The decrease of polyp number and polyp size in the double knockout of Egfr and Erbb3 suggests these two receptors are essential for the development of intestinal polyps. Unlike the small intestinal polyps in the ApcMin/+ model, epithelial-specific deletion of Erbb3 in the AOM model did not result in tumor size reduction, although the number of polyps was increased. This difference could be due in part to the fact that ApcMin/+ mice develop polyps by loss of APC, while in the AOM model, tumors are induced by stabilization of beta-catenin. However, recent gene expression profiling shows that these two models are highly similar [45], suggesting that the difference in the route of tumor initiation in the ApcMin/+ and AOM models likely does not contribute to molecular differences resulting in ERBB3 sensitivity. An alternative possibility is that a subset of intestinal polyps can grow independently of ERBB3, similar to previous results observed for EGFR [40,46] In this study, we observed a profound genetic background-dependent effect on tumorigenesis in the absence of ERBB3. The effect on tumorigenesis of removing ERBB3 may result from its unique link to PI3K/AKT and its downstream effector MAPK. Furthermore, as ERBB3 partners with other ERBB receptor and triggers essential downstream signals, a lack of ERBB3 would abolish EGFR/ERBB3, ERBB2/3, and ERBB3/4 heterodimers simultaneously, which may contribute to its impact on polyp size. The decreased transcript levels of ERBB4 previously reported in ERBB3-deficient intestinal tumors suggests that elevated apoptosis may be due to loss of ERBB3-ERBB4 heterodimers [23], consistent with experiments showing a requirement for ERBB3-ERBB4 heterodimer–dependent AKT pathway activation to prevent CRC cell apoptosis [23]. Consequently, targeting ERBB3 and disrupting heterodimer formation, or using antibodies that inhibit ERBB3 heterodimerization with other ERBBs, may be more efficient than targeting individual receptors, but with the consequence of increased intestinal proliferation and polyp formation in some patients. Our study highlights the importance of regulators of intestinal tumor progression that are dependent on the ERBB3 signaling pathway. It will be important to determine whether ERBB3-dependent signaling also contributes in a genetic background-dependent manner to tumorigenesis in other cancers such as breast, lung, and prostate cancers, where PI3K/AKT is also strongly implicated. These results have important clinical implications that could improve precision applications ERBB3 inhibitor therapy.

Materials and methods Ethics statement All animal procedures were approved by the Institutional Animal Care and Use Committee at Texas A&M University conforming to the Guide for the Care and Use of Laboratory Animals. Animal experiments All animal studies were maintained, and protocols followed, in accordance with Texas A&M University Institution Animal Care and Use Committee guidelines. Mice were obtained from The Jackson Laboratory (C57BL/6J (B6)-ApcMin/+) and NCI-Frederick (B6;D2-Tg(Vil-Cre)20Syr) and maintained hemizygous on the C57BL/6J background. Tg(Vil1-Cre) was selected since it results in extensive target gene deletion throughout the intestinal and colonic epithelia with little expression elsewhere and was used in previous studies with EGFR and ERBB3 [23,41]. Erbb3tm1.1Dwt (Erbb3f) and Egfrtm1Dmt (Egfrf) mice were maintained on the C57BL/6J background. All alleles were also crossed to the 129S1/SvImJ background and intercrossed to make the B6129 hybrid background. The B6;129 mixed background was reproduced by generating F3-F5 advanced intercrosses. Mice were housed five per cage, fed Purina Mills Lab Diet 2919, and maintained at 22° under a 12-hr light cycle. Mice were euthanized by CO 2 asphyxiation for tissue collection. Genotyping Mice were genotyped for the ApcMin allele as previously described [40]. Cre transgenic mice were identified using PCR with cre-S1, 5-gtgatgaggttcgcaagaac and cre-AS1, 5-agcattgctgtcacttggtc primers producing a 278-bp PCR product. Mice were genotyped for the Egfrtm1Dmt allele as previously described [47] and for the Erbb3tm1.1Dwt allele using B3-F, 5’- TCCAGCGTGGAAAAGTTCAC; and B3-R, 5’- AAGCCTTCTCTATGGAAAGTG. Azoxymethane (AOM) carcinogenesis A single lot of AOM was obtained from Sigma-Aldrich (St. Louis, MO) in 100 mg isovials and stored at -80°C. Each 100 mg vial was resuspended in 2 ml phosphate-buffered saline (PBS) and individual 250 ul aliquots stored at -80°C until use. A working stock of 1.25 mg/ml AOM was made by diluting individual 250 ul aliquots into 10 ml of saline (0.9% NaCl). Three-month-old mice were injected intraperitoneal (IP) 10 mg AOM per kg body weight once a week for 4 weeks as described [48,49]. Age-matched controls were injected with saline. Tissue collection The small intestine and colon were removed at three months of age from ApcMin/+ mice or colons five months after last carcinogen dose from AOM-treated mice. Small intestine was cut into quarters, and each segment was gently flushed with PBS to remove fecal material, cut longitudinally, and splayed flat. Representative tumors were scored before bisecting under a dissecting microscope. One half was used for molecular analysis; the other half was fixed for histological analysis, or snap-frozen for use in cryo-sectioning. Macroadenoma counts Tumor numbers were counted and diameters measured along the entire length of the small intestine and colon with a dissecting microscope and in-scope micrometer at 5x magnification without knowledge of genotype by the investigator. Tumors approximately 0.3 mm in diameter were the smallest that could be counted. Changes in tumor growth rate across genotypes were estimated based on tumor size at date of euthanasia. Tumors were also scored based on location along the intestinal tract. Averages and standard deviations for all polyp counts and sizes are provided (S1 Table) as well as representative whole-mount images (S3 Fig). Approximately 5% of representative polyps were verified by histology. Histology and immunohistochemistry Samples from the ApcMin/+ model were fixed in 10% neutral buffered formalin at 4°C overnight before embedded in paraffin and collecting 7 μm sections every 50 μm for staining. Immunohistochemical procedures were performed as described [50]. Colon tumors were dissected, fixed in 4% paraformaldehyde, and embedded in paraffin before cutting 10 μm sections. After antigen-retrieval by boiling for 20 min in citrate buffer, pH 6.0, sections were treated with 0.3% hydrogen peroxide in PBS for 30 min, washed in PBS, blocked in 3% specific serum and 0.1% Triton X- 100 in PBS, and then incubated with primary antibodies and HRP-conjugated specific anti-rabbit secondary antibody (Vector Laboratories, Inc). DAB peroxidase substrate kit (Vector Laboratories, Burlingame, CA) was used to detect antigen-antibody complexes according to the manufacturer’s protocol. Normal and tumor samples were randomly chosen by blinded personnel, and the samples were then confirmed via hematoxylin and eosin staining. Proliferation (Ki67, ab15580-Abcam) and apoptosis (TUNNEL Assay, ab206386-Abcam) was assessed on five samples of each genotype (with and without ERBB3) on the C57BL/6J background. Cells were scored as positive when nuclear staining was evident. Normal tissue was defined as areas where crypts were well orientated. In tumor regions, the total number of cells and all positively stained cells in the core of the tumor were counted using Fiji ImageJ2 [51] and expressed as a percentage of the total number of cells counted. For EGFR IHC, 5 μM sections were taken and antigen retrieval performed by boiling in sodium citrate buffer pH 6.0 for 20 minutes. Sections were blocked in TBS plus 10% normal goat serum and 1% BSA for 1 hour before being incubated at 4°C overnight with primary antibody (1:100, ab5652-Abcam) diluted in TBS plus 1% BSA. Sections were then treated with 0.3% hydrogen peroxide in TBS for 15 minutes before incubation with HRP-conjugated secondary antibody (1:2000, ab205718-Abcam) diluted in TBS plus 1% BSA for 1 hour. Antigen-antibody complexes were detected with ImPACT DAB peroxidase substrate (Vector Laboratories SK4105) according to the manufacturer’s protocol. Tissues were counterstained with hematoxylin for 5 minutes before imaging. Transcriptomic analysis Three sequencing runs were performed to sequence 56 samples on a NextSeq500 by the Texas A&M Institute for Genome Sciences and Society using high output kit v2. A total of 1.5 billion 75 bp single-end reads were checked for adapter sequences and low-quality bases using Trimmomatic [52], resulting in approximately 1.4 billion filtered reads (96%). RNA-Seq reads were aligned to mouse assembly mm10 using HISAT2 version 2.0.5 [53] with an overall mapping rate of approximately 97%. Raw gene counts were generated with feature Counts package [54], while discarding ambiguous read mappings. Normalized read counts and gene expression tests were performed using DESeq2 following recommended guidelines [55]. Ingenuity Pathway Analysis (IPA) was used to analyze differentially expressed genes between groups. All RNAseq data is available at NCBI BioProject ID PRJNA635118. Quantitative real time PCR (qRT-PCR) Egfr in normal and tumor tissues from C57BL/6J and B6;129 mixed background mice were analyzed by qPCR as previously [41]. Genes with significant changes in expression between polyps with and without Erbb3 were identified using ANOVA and select genes confirmed by qRT-PCR. cDNA was synthesized from total RNA from each tumor using the QuantiTect Reverse Transcription Kit (Qiagen 205314). PCR reactions were set up in 96-well plates, and all samples were run in triplicate. Analysis was performed on a LightCycler 96 Thermocycler (Roche) using LightCycler 480 Sybr Green I Master reaction mix. Specific primers were designed to amplify a fragment from genes implicated in the RNAseq analysis (S2 Table). Statistics Nonparametric Mann–Whitney U test was used to analyze polyp data. To compare the statistical difference between two groups, student’s t test was used. P-values smaller than 0.05 were considered significant.

Acknowledgments The authors thank Mr. Kranti Konganti for assistance with RNAseq analysis.

[END]

[1] Url: https://journals.plos.org/plosgenetics/article?id=10.1371/journal.pgen.1009931

(C) Plos One. "Accelerating the publication of peer-reviewed science."
Licensed under Creative Commons Attribution (CC BY 4.0)
URL: https://creativecommons.org/licenses/by/4.0/


via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/