CD142 Identifies Neoplastic Desmoid Tumor Cells, Uncovering Interactions Between Neoplastic and Stromal Cells That Drive Proliferation

The interaction between neoplastic and stromal cells within a tumor mass plays an important role in cancer biology. However, it is challenging to distinguish between tumor and stromal cells in mesenchymal tumors because lineage-specific cell surface markers typically used in other cancers do not distinguish between the different cell subpopulations. Desmoid tumors consist of mesenchymal fibroblast-like cells driven by mutations stabilizing beta-catenin. Here we aimed to identify surface markers that can distinguish mutant cells from stromal cells to study tumor–stroma interactions. We analyzed colonies derived from single cells from human desmoid tumors using a high-throughput surface antigen screen, to characterize the mutant and nonmutant cells. We found that CD142 is highly expressed by the mutant cell populations and correlates with beta-catenin activity. CD142-based cell sorting isolated the mutant population from heterogeneous samples, including one where no mutation was previously detected by traditional Sanger sequencing. We then studied the secretome of mutant and nonmutant fibroblastic cells. PTX3 is one stroma-derived secreted factor that increases mutant cell proliferation via STAT6 activation. These data demonstrate a sensitive method to quantify and distinguish neoplastic from stromal cells in mesenchymal tumors. It identifies proteins secreted by nonmutant cells that regulate mutant cell proliferation that could be therapeutically. Significance: Distinguishing between neoplastic (tumor) and non-neoplastic (stromal) cells within mesenchymal tumors is particularly challenging, because lineage-specific cell surface markers typically used in other cancers do not differentiate between the different cell subpopulations. Here, we developed a strategy combining clonal expansion with surface proteome profiling to identify markers for quantifying and isolating mutant and nonmutant cell subpopulations in desmoid tumors, and to study their interactions via soluble factors.


Introduction
Stromal cells of the tumor microenvironment play an important role in disease progression and therapeutic response. However, the investigation of their these tumors do in fact contain a subpopulation of mutant cells but at too low a frequency to be detected by Sanger sequencing (5,6). Despite the use of cryomolds to dissect contaminating tissues in these next-generation sequencing efforts, CTNNB mutation was only detectable in a minority of the reads per sample (5). We previously found that desmoid tumors can be derived from Ng2-expressing perivascular cells in mice (12). In that investigation, we also found a population of cells in the murine tumors which did not derive from these tumor-initiating cells. Taken together, this suggests that desmoid tumors exhibit strong intratumoral heterogeneity. Some cells may be non-neoplastic stromal cells infiltrating the tumor mass, which makes Sanger sequencing not suitable to detect CTNNB mutations in a subset of cases.
Tumor-associated fibroblasts predominate in the cellular tumor microenvironment and their role in disease progression is well documented (13,14). These mesenchymal cells are implicated in regulating the neoplastic phenotype of several tumors including breast cancer (15,16), squamous cell carcinoma (17), non-small cell lung cancer (18), and scirrhous gastric carcinoma (19) among many others. Tumor-associated fibroblasts can stimulate tumor progression and invasiveness (14,17,20), but they can also restrain tumor progression in other contexts (21)(22)(23). Recently, they have also been implicated in therapeutic response and drug resistance (24,25). Little is known about the tumor stroma and its effect on neoplastic cells in mesenchymal tumors. This is in part because investigating the role of the tumor stroma in mesenchymal lesions has been challenging because the neoplastic cells exhibit the same mesenchymal phenotype as the stromal fibroblasts. In fibroproliferative diseases like desmoid tumors, the distinction is doubly challenging as both groups exhibit a fibroblastic phenotype. Surface proteins that differ in their expression between the two groups offer a potential identifying feature. While cell surface profiling of fibroblasts has been conducted previously (26), how the neoplastic and non-neoplastic mesenchymal cells within desmoid tumors compare is not known. We therefore sought to study the expression of surface proteins in single cell-derived samples of desmoid tumors.
Neoplastic and stromal cells can communicate via paracrine signaling (27). This cross-talk can modulate tumor initiation, progression, and invasion (13,14). Previous studies of tumor-stroma interactions have identified a number of important pathways in tumor development, including TGFβ, NFκB, PDGF, Notch, and Src (14,20). Variability in behavior of other tumors caused by beta-catenin activation, such as colorectal cancers, is linked to the tumor microenvironment playing a major role in modulating beta-catenin activity and tumor progression (28). While it is well known that cytokines such as TGFβ stimulate fibroblast proliferation, mediated by beta-catenin (29), the role of cell-cell communication via paracrine signaling in beta-catenin-driven desmoid tumors has not been investigated. In this study, we isolated neoplastic and stromal cells using clonal expansion. A high-throughput surface antigen screen was used to analyze the expression of proteins on the cell surface. We used these cells to identify and study cell-cell interaction by paracrine signaling and identified a possible therapeutic target based on disrupting this interaction.

Human Desmoid Tumor Cell Cultures and Generation of Monoclonal Cell Populations
Human desmoid tumor tissue samples were obtained at time of surgical excision and processed for primary cell culture generation as described previously (30,31) followed by long-term storage in liquid nitrogen. The use of human material was approved by the Institutional Review Board of the Hospital for Sick Children. Because each primary culture exhibited a different colony-forming capacity, we initially performed limiting dilutions (within a range of 1 to 20 cells per 1 well of a 24-well plate) to determine the best concentration to establish single cell-derived colonies for each line. Each well was visually inspected for the presence of a single colony. A cluster of 32 or more cells at 8 days after seeding was counted as a colony. Cells were maintained at 37°C in 5% CO 2 in normal culture media consisting of DMEM (Wisent, 319-005-CL, containing 4.5 g/L glucose, with l-glutamine, sodium pyruvate, and phenol red), supplemented with 10% FBS (Wisent, 080-450) and antibiotic/antimycotic solution (Wisent, 450-115-EL). For monoclonal cultures, the media was changed daily. For primary cultures, media was changed three times a week, and the cells were passaged at a 1:3 ratio when confluent. All primary cultures for single cell-derived colony formation were used at or before the third passage. For all other experiments, cells were utilized at under six passages. Cells not immediately used for downstream experiments were stored in liquid nitrogen in freezing media (normal culture media with 10% DMSO).

Sequencing Analysis
Whole-genomic DNA was extracted from approximately 1 × 10 5 primary and approximately 1 × 10 4 clonal expansion cells using DNeasy Blood & Tissue Kit (Qiagen, 69506) as per the manufacturer's instructions with some modifications. Briefly, the cells were pelleted by centrifugation at 800 × g for 3 minutes and resuspended in 200 μL PBS. To elute the DNA, 75 and 40 μL of water was added to the spin column containing primary and monoclonal samples, respectively, and left to incubate at room temperature for 4 minutes before elution. PCR amplification of exon 3 of CTNNB was conducted using 100 ng of extracted genomic DNA and the Q5 High-Fidelity Master Mix (New England Biolabs, M0492L) in a 50 μL reaction containing primers designed to amplify the human CTNNB exon 3 region (Forward Primer: TCCAATCTACTAATGCTAATACTGTTTCGTA, Reverse Primer: CATTCT-GACTTTCAGTAAGGCAATG). PCR amplification was carried using 2720 Thermal Cycler (Applied Biosystems) with the following program: 98°C for 30 seconds initial denaturation. A total of 30 cycles of 98°C for 10 seconds, 60°C for 30 seconds, and 72°C for 20 seconds. Final extension was carried at 72°C for 2 minutes and the product was held at 4°C. The PCR product was purified by running it through 1.7% agarose gel and using the QIAquick Gel Extraction Kit (Qiagen, 28706) as per the manufacturer's instructions with some modifications. Briefly, the column was incubated with Buffer PE for 5 minutes at room temperature before centrifugation. To elute, the DNA was incubated with 30 μL water for 4 minutes at room temperature. Approximately 150 ng of the PCR product was analyzed by Sanger sequencing (TCAG, Hospital for Sick Children). Initial inspection of the chromatograms using FinchTV (https://digitalworldbiology.com/FinchTV) stratified samples based on the altered nucleotide. All T41A samples displayed an A>G mutation at codon 41, and all S45F samples displayed a C>T mutation at codon 45. Codons 32, 33, and 37 were also inspected but no sample harbored a mutation at these codons. To estimate the mutation frequency, the peak height data of the sequencing chromatograms were then analyzed using QSVanalyzer (32)

High-throughput Surface Antigen Screen
Cells were analyzed by a flow cytometry-based cell surface antigen screen as described previously (26,33). Cells were resuspended in 10 mL of PBS, with 2 mmol/L ethylenediaminetetraacetic acid (EDTA) and 2% FBS, and were passed through a 70 μm cell strainer to reduce cell clumping. Colonies of the same genotype were pooled into one cell suspension. Cells from mutant and nonmutant colonies were aliquoted over multiwell plates each containing a unique fluorochrome-tagged antibody targeting one known cell-surface antigen, for a total of 368 antibodies. Gating strategy was conducted as described previously (26,33). A representative gating example is shown in Supplementary  Fig. S12A-S12D. The experiment was repeated twice, once for each mutant subgroup (T41A and S45F). A "hit" in each experiment was considered an antigen with a detectable expression in greater than 50% of the cells in one population and less than 10% in the other. Alternatively, an antigen expressed by more than 50% of cells in one population and differed by more than 10-fold in median fluorescent intensity was also considered as a hit. Surface proteins that were detected as hits in both experimental repeats were considered candidate markers for further study.

Surface Marker-based FACS
Cells were resuspended in sterile cold staining buffer (0.1% gelatin, 2 mmol/L EDTA in PBS) at a concentration of 1 × 10 7 cells/mL. PE Mouse anti-human CD142 Clone HTF-1 (BD, 550312), PE Mouse Anti-Human OX40 Ligand (CD252) Clone ik-1 (BD Biosciences, catalog no. 558164, RRID:AB_647195), or Alexa Fluor 647 anti-human Podoplanin Antibody Clone NC-08 (BioLegend, catalog no. 337008, RRID:AB_2162063) were then applied to cell suspension at a 1:20 dilution factor and the antibody-cell mixture was incubated in the dark at 4°C for 45 minutes. The mixture was diluted 5-fold with cold staining buffer and centrifuged at 800 × g for 3 minutes. The cell pellet was then resuspended at 1 × 10 7 cells/mL with staining buffer and kept on ice until cell sorting. Propidium iodide (Thermo Fisher Scientific, P3566) was used to detect dead cells. Representative gating strategy is shown in Supplementary Fig. S12E.

Conditioned Media Generation and Secretome Profiling
A total of 1.5 × 10 5 cells were plated in each well of a 6-well plate in normal media. After 24 hours, the media was changed to DMEM containing 1% FBS and the cells were incubated for a further 48 hours. The media was subsequently collected and centrifuged at 800 × g for 3 minutes at 4°C. The supernatant was transferred to new tubes and stored at −80°C. Secreted protein detection of 105 different cytokines was carried out using the Human XL Cytokine Array Kit (R&D, ARY022B) as per the manufacturer's instructions. Imaging was conducted using the ChemiDoc MP Imaging System (Bio-Rad) and spot pixel density measurements were performed using ImageJ (RRID:SCR_003070) via the Protein Array Analyzer module. For phosphorylated kinase detection after conditioned media treatment, we carried out phosphokinase detection of 43 proteins with Proteome Profiler Human Phospho-Kinase Array (R&D, ARY003B) per the manufacturer's instructions.

Proliferation Assays
For coculture studies, 1.5 × 10 5 cells were seeded onto sterile glass cover slips in the bottom chamber of a well in a 6-well plate divided by an 8.0 μm pore size Transwell permeable support (Corning, 3428). Twenty-four hours before staining, the BrdU reagent (Invitrogen) was added to the culture media at a 1:100 dilution. The following day, the cells were fixed with ice-cold methanol for 15 minutes at −20°C. The cells were then washed with PBS for 3 minutes, repeated for a total of three times. Permeablization was carried out using the FixDenat solution (Roche, 11647229001) for 30 minutes at room temperature. The cells were then washed with PBS three times and blocked using a blocking solution of 2% horse serum, 2% donkey serum, and 2% BSA in PBS-  Table S5 lists the primers used in these experiments. Relative gene expression levels were calculated using the C t method.

Immunoblotting
Protein lysates were generated using Reporter Lysis Buffer (Promega, E397A). Approximately 10 μg of protein was boiled for 10 minutes in Laemmli buffer with 5% β-mercaptoethanol and 2% SDS. The reduced proteins were resolved on a 7.8% polyacrylamide gel. After transfer to a nitrocellulose membrane, the membranes were blocked with 5% BSA in TBS-T solution followed by overnight incubation at 4°C with primary antibodies against

Immunofluorescence
Serial sections of a formalin-fixed and paraffin-embedded tumor biopsy with low and high beta-catenin areas were used for immunofluorescence. After rehydration, heat-mediated antigen retrieval was performed by heating the sections in citrate buffer (pH 6.0) in a pressure cooker. The sections were then blocked with 1% BSA + 10% serum (matching secondary antibody host) in PBS-

Recombinant Protein and STAT6 Inhibitor Treatment
All recombinant proteins were purchased from R&D Systems, reconstituted as per the manufacturer's instructions and used at the following concentrations:

Statistical Analyses
Statistical calculations were conducted using R (version 4.2.2). k-nearest neighbor distances were calculated using kNNdist function from the dbscan package (version 1.1-11). Statistical significance was evaluated using the one-and twosample t tests from base R, with a P value less than 0.05 considered significant.

Data Availability
The data generated in this study are available within the article and its Supplementary Data.  Fig. 1). From this analysis, we found that the estimated proportion of mutant cells ranged from approximately 20% to 95% (mutation frequency for all samples was 54% on average, with a standard deviation of 27%; Table 1). There was no appreciable difference in the estimated average mutation frequency between the T41A and S45F samples (49% and 66%, respectively, unpaired t test P = 0.31). No sample with a detectable mutation displayed more than one mutational event. We also observed that the mutation frequency declines with subsequent passages in vitro (Supplementary Fig. S1), suggesting that the current culture conditions favor nonmutant cell growth. Thus, subsequent experiments utilizing primary cultures were conducted on low-passage cultures only (maximum of five passages), and samples were resequenced to confirm their mutational status. Our results demonstrate high variability in the cellular composition of desmoid tumor primary cultures.  Table 1.

Desmoid Tumors Contain CTNNB1 Mutant and Nonmutant Mesenchymal Cells
To identify the mutant and nonmutant cells in heterogeneous cultures, we established single cell-derived colonies from two primary cultures (one T41A, and one S45F), both displaying medium (40%-60%) CTNNB mutation frequency. While cells from all colonies displayed a general spindle-like shape, we observed that some colonies exhibited provisionally distinct morphologic and spatial patterns: we observed colonies whose cells tend to appear condensed in morphology and tend to cluster together (we labeled these colonies as Type A), and colonies whose cells display generally more flat morphology and tend to spread apart (labeled as Type B; Supplementary Fig. S2A and S2B).
This phenotype was persistent upon subsequent subcloning suggesting a heterogeneity of fibroblastic cell types within the tumor mass. These phenotypic differences resembled those previously reported on fibroblastic monocultures (34,35). Specifically, our observed Type A colonies resembled fibroblasts while Type B colonies resembled myofibroblasts. Consistent with a myofibroblast-like phenotype, Type B cells expressed higher smooth muscle alpha-actin compared with Type A both at the RNA and protein level (Supplementary Fig. S3).
We then analyzed at least three colonies for each subphenotype by Sanger sequencing of CTNNB. Peak height analysis of sequencing chromatograms supported the clonality of each subpopulation derived using this approach.
Type A colonies with clustered growth showed no detectable mutations in CTNNB, while Type B colonies with dispersed growth carried the mutation detected in their original primary culture ( Fig. 2A). Because the detected betacatenin mutations are known to increase its transcriptional activity (4, 31), we analyzed the gene expression of AXIN, a known transcriptional target of beta-catenin (36). As expected, the mutant colonies exhibited elevated AXIN expression compared with nonmutant colonies (Fig. 2B). We detected no difference in AXIN expression between T41A and S45F mutant colonies. We also established additional single cell-derived colonies from two high (>80%) and three low (<40%) mutational frequency primary cultures but we these colonies did not exhibit distinct morphologic or genotype differences.

High-throughput Cell Surface Profiling Distinguished Mutant from Nonmutant Colonies in Desmoid Tumors
Proteins anchored to the cell membrane can be used as relatively accessible markers to distinguish between cell populations, but such proteins have not been identified in mesenchymal tumors. We performed a high-throughput surface antigen screen on two desmoid tumor samples (one T41A and one S45F), as described previously (26,33), to evaluate the expression of 368 surface antigens by flow cytometry to identify such markers (Supplementary Table S1).
In both samples, we profiled CTNNB mutant colonies and CTNNB wildtype colonies derived from the same desmoid sample, for a total of four screens. In general, we observed a very similar surface protein expression profile between mutant and wildtype colonies. This profile was consistent with previously reported fibroblast surface proteome. We found 24 proteins that were present in more than 50% of cells within all tested colonies (whether CTNNB mutant or not). This class of proteins included known fibroblastic markers such as CD13, CD29, CD44, and CD73 (37-39). When we compared our results with a previously published cell surface profile of normal dermal fibroblasts using the same flow cytometry-based screen (26), we found 21 (88%) of the 24 expressed proteins were also positive in dermal fibroblasts; the remaining three proteins were still detectable in more than 10% of normal dermal fibroblasts. Conversely, 290 proteins were detected in less than 50% of cells in all tested subgroups in our study. Of these, we found 278 (96%) proteins were also negative in the dermal fibroblast profile; four proteins in our screen were expressed in more than 50% of normal dermal fibroblasts (CD21, CD66e, CD104, and CD151), and the remaining nine were not tested in the previously published profile. Our results are therefore comparable with previously published work and provide additional evidence for the reproducibility of this screen despite the small number of biological replicates. Given the fibroblastic nature of desmoid cells, we expectedly found that mutant and nonmutant colonies shared a very similar fibroblastic expression pattern for most proteins probed in this study. For example, five proteins (CD49E, CD90, CD73, CD44, and CD13) were previously identified as candidate cancer-associated fibroblast markers in serous ovarian cancer using this screening strategy (26), but here we found all proteins were expressed by all subpopulations at greater than 80% positivity (Supplementary Table S1), supporting the need to profile distinct subpopulations within mesenchymal tumors.
Despite the similarity in the cell surface profiles, few surface proteins tested were expressed primarily by one subgroup (Supplementary Table S2). Of these, CD142 (also known as tissue factor, TF, encoded by F) was identified as a candidate positive marker of mutant cells in both experiments (Fig. 3A).
We also identified podoplanin as a marker enriched in the nonmutant population ( Supplementary Fig. S4). To validate the results, we compared the relative gene expression level of CD142 (encoded by F) and podoplanin in a larger set of mutant and nonmutant colonies derived from six samples. We observed that the mutant populations have higher expression of F relative to nonmutant populations derived from the same desmoid tumor samples (n = 6, P < 0.05; Fig. 3B). Conversely, podoplanin expression is lower in the mutant colonies. We next aimed to study the expression of our candidate surface marker in situ. First, we performed double immunofluorescent staining for CD142 and PDPN in a human desmoid tumor biopsy showing high and low beta-catenin staining regions (Fig. 3C). We observed that CD142 staining matched beta-catenin staining pattern. Conversely, podoplanin staining was higher in low beta-catenin regions. To determine whether CD142 is expressed in a broader range of desmoids, we performed IHC staining on a tissue microarray  (Table 2). We then examined dermal scar tissues, since we have previously shown that beta-catenin is stabilized in these tissues (29,40), and detected positive CD142 staining in a panel of 10 core biopsies from scar tissues (Supplementary Table S3; Supplementary Fig. S11). Thus, CD142 overexpression may be a marker of beta-catenin activation, and not restricted to CTNNB mutational status. Treatment of wildtype dermal fibroblasts with recombinant Wnt3a showed a trend toward increased CD142 expression, encoded by F, but the differences were not statistically significant ( Supplementary Fig.   S5). Taken together, our data identify CD142 and podoplanin as potential surface markers that are differentially expressed between cells with mutant, active beta-catenin and nonmutant, inactive beta-catenin in heterogeneous desmoid tumor samples.

CD142 Identifies the Mutant Cell Population in Desmoid Tumors
To investigate the utility of CD142 to isolate mutant cells from heterogeneous primary desmoid tumor cultures, we conducted CD142-based flow cytometry and FACS of three primary cultures with a known mutation and the two primary cultures in which we did not detect any mutations in CTNNB using Sanger sequencing. Analysis of these samples shows the proportion of CD142-positive (CD142 + ) cells correlated with the estimated mutation frequency from Sanger sequencing experiments conducted in parallel (Pearson correlation coefficient r = 0.9874, P = 0.0017; Fig. 4A; Supplementary Fig. S6).
We then expanded the sorted CD142 + cells and analyzed them by Sanger sequencing. In the three primary cultures with previously known mutation, the and decreased podoplanin (PDPN) in mutant colonies (n = 6 samples, * one-sample t test P < 0.05). C, Immunofluorescence staining of CD142, PDPN and beta-catenin in a desmoid tumor sample (T: Tumor, S: Stroma, as identified by visually inspecting beta-catenin staining intensity).
CD142 + cells were enriched (>95% estimated mutation frequency) for the same CTNNB mutation. Notably, 8.67% of the cells in one of the two primary cultures previously labeled as "wildtype" were also labeled as CD142 + . Subsequent expansion and sequencing of these cells revealed an S45F mutation at approximately 50% mutation frequency. Our detection of a minor mutant subpopulation in one of the samples that failed mutation detection by Sanger is consistent with findings that traditional sequencing methods of analyzing bulk tumors are not sensitive enough to detect a small population of mutant cells (5,6). We then studied the utility of multicolor staining in our FACS experiments. In CD142/PDPN double-staining experiments, we detected both CD142 High ;PDPN Low and CD142 Low ;PDPN High populations which corresponded to subpopulations enriched for mutant and nonmutant CTNNB, respectively ( Fig. 4B; Supplementary Table S4). FACS isolated mutant and nonmutant subpopulations displayed similar morphologic differences in monoculture ( Supplementary Fig. S7) as was previously observed when comparing single cell-derived mutant and nonmutant colonies. Over-all, CD142-based FACS offers an alternative method to enrich for the mutant population in heterogeneous desmoid tumor primary cultures.

Stroma-derived Secreted Factors Induce Mutant Cell Proliferation
Stroma-derived soluble factors that function in paracrine signaling can alter tumor cell behavior. To determine whether paracrine signaling is active in desmoid tumors, we cocultured mutant desmoid tumor cells with nonmutant fibroblasts using permeable inserts to allow soluble factors to be shared between the two populations without direct cell contact. CCL2, CXCL12, and CHI3L1 in serum-free conditions. We observed that PTX3, CCL2, and CXCL12 significantly increased BrdU incorporation compared with no treatment (Fig. 5D). Therefore, our data demonstrate differential expression of secreted factors in mutant and nonmutant cells in desmoid tumors. Moreover, nonmutant fibroblasts have the potential to increase the proliferation of mutant desmoid tumor cells by secreting soluble factors that include PTX3, CCL2, and CXCL12.

The STAT6 Inhibitor AS1517499 Blocks Stroma-induced Proliferation of Mutant Cells
To investigate the pathways activated in by the conditioned media, we used a phosphokinase antibody array to analyze the mutant cells treated with nonmutant conditioned media and verified results by Western blotting. We detected increased phosphorylation of STAT6 following conditioned media treatment ( Fig. 6A and B). To determine whether the identified stroma-derived secreted factors contributed to STAT6 phosphorylation, we treated mutant desmoid tumor cells with recombinant secreted factors differentially expressed by the nonmutant cells and found that PTX3 induced the highest increase in phospho-STAT6 (Fig. 6C). To test whether STAT6 activity influenced mutant cell proliferation, we used AS1517499, a potent STAT6 inhibitor (41).
Treating mutant cells with nonmutant conditioned media in combination with 50 nmol/L of AS1517499 abrogated the effect of increased proliferation from treating with conditioned media alone (Fig. 6D). To determine whether PTX3 induces desmoid tumor cell proliferation via STAT6 activation, we combined recombinant PTX3 protein treatment with AS1517499. We found that while PTX3 treatment alone increased BrdU incorporation as we found previously, STAT6 inhibition by AS1517499 abrogated this increase (Fig. 6E). Taken together, our data suggest that nonmutant stromal fibroblasts induce mutant desmoid tumor cell proliferation via secreting soluble factors, including PTX3 and STAT6 activation (Fig. 6F).

Discussion
Here we describe the isolation of mutant and nonmutant subpopulations from heterogeneous desmoid tumor samples by clonal expansion and surface marker-based cell sorting. Our study identifies CD142 as a marker enriching for the mutant population which can be used to isolate the mutant population using conventional cell sorting methods for downstream profiling. Furthermore, we found that nonmutant cells can contribute to mutant cell proliferation via secreted factors. PTX3, CCL2, and CXCL12 are stroma-derived factors that increase the proliferation rate of mutant desmoid tumor cells. STAT6 partially mediates the ability of stromal cells to induce proliferation in mutant cells.
The mutation status of desmoid tumors has traditionally been determined by Sanger sequencing and not all tumors had mutations identified. Deep sequencing to identify minor mutant populations within heterogeneous samples (5) showed a higher mutation rate. The proportion of mutant and nonmutant cells within the tumor may contribute to differences in mutation detection rates, especially when Sanger sequencing is utilized. Interestingly, there has been a suggestion that nonmutant tumors have a better prognosis (42)(43)(44). It is possible that the "wildtype" tumors in these studies had beta-catenin mutations in a small minority of tumor cells, and instead the better prognosis in "wildtype" tumors could be related to a lower mutant tumor cell burden. An analysis of CD142, also called TF, is a member of the blood coagulation cascade, but recent evidence suggests it may also be involved in additional cellular signaling important for tumor growth (45). Structurally, CD142/TF shares an evolutionary lineage with the type II cytokine receptor family which also includes the IFN receptors and IL receptors. Mechanistically, CD142/TF can activate a signaling pathway by its interaction with FVIIa which has been shown to lead to progrowth ERK1/2 phosphorylation. These mechanisms have been implicated in angiogenesis (46) and metastasis (47)   Circles indicate individual datapoints.
Our study has some limitations. It is possible that surface protein expression is influenced by the specific CTNNB variant. Given the small number of samples in our study, however, we cannot distinguish whether the higher general expression of markers (e.g., CD120A, CD123, CD178) in cells derived from T41A mutant tumors thought to be a result of tumor biology with differences in both tumor and stroma resulting from that specific mutation and subsequent signaling alterations, or artifact of the two experiments being performed separately.
We therefore focused our study on overlapping markers, a criterion only CD142 met in this study. Future experiments that are conducted at a larger scale can help distinguish whether specific CTNNB variants yield a difference in tumor and stroma surface protein expression. It is also possible that additional proteins found in a minority of our cells could still be useful to identify subpopulations which may harbor important biological functions. Moreover, our study is limited by the antibodies included in the high-throughput flow cytometry panel used. As such, the utility of additional surface proteins such as FAP, which was not included in this panel, could not be determined. Given that the screen was conducted on cultured cells, it is also possible that in vitro culture conditions impacted the expression pattern of some markers.
Our study demonstrates that STAT6-activating secreted factors contribute to mutant cell proliferation. Evidence from other cancer types suggests a role for STAT6 activation in tumor progression. High STAT6 expression correlates with worse prognosis in colorectal cancer (51) and prostate cancer (52). STAT6 is also often found constitutively phosphorylated in Hodgkin and Reed-Sternberg cells of Hodgkin lymphoma (53). Treatment with AS1517499 was previously shown to reduce tumor growth and early liver metastasis in an orthotopic 4T1 mammary carcinoma mouse model (54), suggesting that such a strategy may also be promising for desmoid tumors.
To date, the investigation of the role of stromal cells in mesenchymal tumors is hampered by difficulties distinguishing them from the neoplastic cells. Our study shows that lineage-specific surface markers would perform poorly in distinguishing functionally and molecularly distinct cell subpopulations in a mesenchymal tumor. Here, we identified surface markers to quantify and isolate mutant and nonmutant cell subpopulations in desmoid tumors, and to study their interactions via soluble factors. Despite the relative uniformity in beta-catenin aberrations across the majority of patients with desmoid tumor, https://doi.org/10.1158/2767-9764.CRC-22-0403 | CANCER RESEARCH COMMUNICATIONS CD142 is a Lineage-specific Marker in Desmoid Tumors the disease exhibits a diversity of clinical behavior. Because our study demonstrates the cross-talk between subpopulations, it is possible that the proportions of mutant and nonmutant subpopulations in each tumor, and their respective secretome, may be an important factor that determines disease progression and response to therapy, which warrants further investigation. We also anticipate our clonal expansion and profiling strategy to be useful to study tumor-stroma interactions in other sarcomas. In desmoid tumors, CTNNB hotspot mutations can be sequenced with a single primer pair, but other techniques such as multiplexed amplicon sequencing, hybridization capture panels, or wholeexome/whole-genome sequencing can be deployed in diseases where there is no single disease-defining mutation.