Molecular and Immunologic Techniques in a Genetically Engineered Mouse Model of Gastrointestinal Stromal Tumor

GIST is the most common sarcoma in humans with an incidence of about 6,000 cases in the United States of America¹. GIST appears to originate from the gastrointestinal pacemaker cells named the interstitial cells of Cajal, and is typically driven by a single mutation in the tyrosine kinase KIT or PDGFRA². Surgery is the mainstay of treatment for GIST and can be curative, but patients with advanced disease may be treated with the tyrosine kinase inhibitor (TKI), imatinib. Since its introduction over 20 years ago, imatinib has transformed the treatment paradigm in GIST, improving the survival in advanced disease from 1 to over 5 years^3,4,5. Unfortunately, imatinib is rarely curative due to acquired KIT mutations, so new treatments are needed for this tumor.

Mouse models are an important research tool in the investigation of novel therapies in cancer. Multiple subcutaneous xenograft and patient-derived xenograft models have been developed and investigated in GIST^6,7. However, immunodeficient mice do not fully represent human GIST since GISTs harbor differential immune profiles depending on their oncogenic mutation, and altering the gastrointestinal tumor microenvironment improves upon the effects of TKI therapy^8,9. The Kit^V558Δ/+ mouse has a heterozygous germline deletion in Kit exon 11, which encodes the juxtamembrane domain, the most commonly mutated site in human GIST¹⁰. Kit^V558Δ/+ mice develop a single cecal GIST with 100% penetrance, and tumors have similar histology, molecular signaling, immune infiltration, and response to therapy as human GIST^8,11,12,13. Here, we describe breeding, treatment, and specimen isolation and processing in Kit^V558Δ/+ mice for use in molecular and immunologic research in GIST.

All mice were housed under pathogen-free conditions at the University of Pennsylvania according to NIH guidelines and with approval of the University of Pennsylvania IACUC. Euthanasia was performed following the University of Pennsylvania Laboratory Animal Resources standard operating procedures.

1. Kit^V558Δ/+ mouse breeding

1. Backcross the Kit^V558Δ/+ mice more than 10 times onto a C57BL/6J background using C57BL/6J mice. To do this, breed the male Kit^V558Δ/+ mice with female C57BL/6J mice at a 1:2 ratio.
   NOTE: The homozygous Kit^{V558Δ/V558Δ} genotype is lethal in utero. It is possible to breed female Kit^V558Δ/+ mice with male C57BL/6J mice, but litter size is about half as that seen from female wildtype C57BL/6J mice. Furthermore, female Kit^V558Δ/+ mice produce limited litters after 4 months of age.
2. Genotype the pups at 7-14 days of age by toe clipping to confirm the presence of the Kit^V558Δ/+ genotype. Use the forward primer: TCTCCTCCAGAAACCCATGTATGAA; reporter 1: CCTCGACAACCTTCCA; reverse primer: TTGCGTCGGGTCTATGTAAACAT; and reporter 2: TCTCCTCGACCTTCCA for genotyping.

2. Kit^V558Δ/+ mouse treatment

1. Age and sex match the Kit^V558Δ/+ mice prior to treatment. Use the age- and sex-matched mice in cohorts as tumors from female Kit^V558Δ/+ mice are larger than those in male mice. Treat the mice at 8-12 weeks old, at which time tumors are established (Figure 1).
2. Give tyrosine kinase inhibitors orally or by intraperitoneal (i.p.) injection; Kit^V558Δ/+ tumors are sensitive to tyrosine kinase inhibitors. Provide imatinib in a dose of 600 mg/L in drinking water or i.p. injections of 45 mg/kg twice daily. Measure tumor weight reduction using digital scales after dissection of tumor as shown in step 3, which is approximately 50% at 1 week and 80% at 4 weeks following treatment with imatinib (Figure 2).

3. Kit^V558Δ/+ mouse organ harvest

1. Euthanize mice by CO₂ narcosis at a flow rate of 60% chamber volume per min. Leave mice in the chamber for at least 2 min after respiration has ceased, then perform cervical dislocation to confirm death.
2. Sterilize all instruments, wear gloves throughout the procedure, and maintain a sterile field. Prepare the skin with 70% ethanol. Make a 2 cm midline vertical incision using scissors and enter the abdominal cavity. Sharply lyse any intra-abdominal adhesions.
3. To remove the draining mesenteric lymph node, follow the steps described below.
   1. Identify the cecum and lift its mesentery superiorly. Approximately midway to the base of the colonic mesentery, identify the mesenteric lymph node and sharply dissect it. The lymph node is off-white and about 0.5 cm x 0.5 cm in size.
   2. Divide lymph node tissue into thirds for protein isolation, histology, and single cell suspension, as needed. For single cell suspensions, place lymph node tissue in 20 mL of serum free media (RPMI) and keep on ice.
4. For isolation of the GIST and cecum, follow the steps described below.
   1. The cecum in Kit^V558Δ/+ mice is mostly replaced by a GIST. Carefully divide the ileocolic junction from the base of the tumor. To collect the cecum, divide the colon again, 2 cm proximal to the base of the tumor.
   2. In 50%-60% of Kit^V558Δ/+ mice, the head of the tumor contains a cap of cecal tissue, which typically contains serous fluid but may rarely contain pus (Figure 3). Sharply dissect the cap tissue away from the tumor tissue.
   3. Divide tumor tissue and/or cecum into thirds for protein isolation, histology, and single cell suspension, as needed. For single cell suspensions, place tumor tissue or cecum in HBSS with 2% FCS, enough to cover the sample and keep on ice.

4. Western blot analysis of GIST tissue

1. Prepare tissue lysis buffer containing 50 mM TrisHCl (pH 7.5), 150 mM NaCl, 5 mM EDTA, 1% nondenaturing detergent, 2 mM Na₃VO₄, 1 mM PMSF, 10 mM NaF, and 20 μl/ml proteinase inhibitor mixture.
2. Prepare a 1x Tris-buffered saline solution with 1% Tween 20 (TBST) by combining 1800 mL of deionized water, 2 mL of Tween 20, and 200 mL of 10x Tris-buffered saline (TBS).
3. Resuspend tissue from step 3.4.3 in a FACS tube in 5mL/g of tissue lysis buffer and homogenize twice with a mechanical homogenizer at 15,000 rpm for 15 s on ice. Incubate lysate for 30 min on ice.
4. Transfer lysate to a 1.5 mL microcentrifuge tube. Centrifuge at max speed for 20 min at 4 °C. Transfer supernatant into a new microcentrifuge tube.
5. Run lysates on a 4% to 15% gradient gel, then transfer to a nitrocellulose membrane, as described in¹⁴.
6. Wash membrane once in 1x TBS for 5 min and then block membrane in 5% milk for 1 h. Again, wash membrane once in 1x TBS for 10 min.
7. Incubate membrane in primary antibody diluted 1:1000 in 5% BSA at 4 °C overnight. Wash membrane 3x in 1x TBST for 10 min each. Incubate blot with secondary antibody diluted 1:2500 in 2.5% milk for 1 h at room temperature.
8. Wash membrane once in 1x TBS for 5 min. Add enough HRP substrate to cover membrane, typically 200-500 μL, and use a digital imager to detect and quantify chemiluminescence.

5. Immunohistochemistry of GIST tissue

1. Fix tissue from step 3.3.2 or 3.4.3 in 4% paraformaldehyde at 4 °C overnight. Store tissue in 70% EtOH until ready for processing. Embed and section blocks at 5 μm thickness onto glass slides, as described^15,16.
2. Complete immunohistochemical detection using a basic immunodetection kit, as previously described¹⁷.

6. Single cell suspension of mesenteric lymph node

1. Pour RPMI media with lymph node specimen from step 3.3.2 over a 100 μm filter. Move filter to new 50 mL conical and mash lymph node with the soft end of a 3 mL plastic syringe. Wash filter with 20 mL of RPMI media.
2. Centrifuge the filtrate at 450 x g at 4 °C for 5 min. Aspirate the supernatant. 
3. Resuspend pellet in 20 mL of 1% FBS in PBS (bead buffer) and pour over a 40 μm filter. Collect the cell filtrate. Count cells using a hemocytometer.
4. Centrifuge the filtrate at 450 x g at 4 °C for 5 min. Aspirate the supernatant. Resuspend in bead buffer at 6 x 10⁷ cells/mL for flow cytometry.

7. Single cell suspension of GIST

1. Prepare collagenase buffer by adding 250 mg of collagenase IV, one tablet of EDTA-free protease inhibitor and 100 µL of DNase I to 50 mL of HBSS. Rotate for 10 min at room temperature until dissolved.
2. Place GIST in a sterile dish and add 2.5 mL of collagenase buffer. Mince tumor using a sterile scalpel and scissors until the tumor is in fine fragments. Use a large bore pipette to aspirate the tumor and collagenase into a 50 mL tube.
3. Incubate in a shaking incubator at 100 rpm at 37 °C for 30 min. Quench reaction with 2 mL of FBS.
4. Pour collagenase with GIST specimen over a 100 μm filter and mash tumor with the soft end of a 3 mL plastic syringe and collect in a 50 mL tube. Wash filter with 20 mL of HBSS. Centrifuge the filtrate at 450 x g, 4 °C for 5 min. Aspirate the supernatant.
5. Resuspend the pellet in 20 mL of bead buffer and pour over a 40 μm filter. Collect the filtrate and count cells using a hemocytometer. Centrifuge the filtrate at 450 x g, 4 °C for 5 min. Aspirate the supernatant. Resuspend in bead buffer at 6 x 10⁷ cells/mL for flow cytometry.

8. Single cell suspension of cecum

1. Prepare collagenase buffer as in step 7.1. Using scissors, split the cecum longitudinally to expose the inner mucosa. Cut into 0.5 cm sections and place in a 50 mL tube with 5 mL of HBSS with 2% FBS. Shake vigorously for 30 s, centrifuge at 450 x g for 20 s, then aspirate the supernatant.
2. Add 5 mL of HBSS with 2 mM EDTA. Incubate in a shaking incubator at 37 °C at 100 rpm for 15 min. Centrifuge at 450 x g for 20 s. Aspirate the supernatant.
3. Add 5 mL of HBSS. Shake vigorously for 30 s, centrifuge at 450 x g for 20 s, then aspirate the supernatant. Repeat once more.
4. Add 5 mL of collagenase buffer. Incubate in a shaking incubator at 37 °C for 30 min at 100 rpm. Shake vigorously every 10 min. Quench reaction with 2 mL of FBS.
5. Pour collagenase with cecum specimen over a 100 μm filter and mash with the soft end of a 3 mL plastic syringe. Wash filter with 20 mL of HBSS. Collect filtrate.
6. Centrifuge the filtrate at 450 x g for 5 min at 4 °C. Aspirate the supernatant. Resuspend pellet in 20 mL of bead buffer and pour over a 40 μm filter. Collect filtrate and count cells using a hemocytometer.
7. Centrifuge the filtrate at 450 x g for 5 min at 4 °C. Aspirate the supernatant. Resuspend in bead buffer at 6 x 10⁷ cells/mL for flow cytometry.

The Kit^V558Δ/+ mouse model allows for the investigation of therapeutics in an immunocompetent mouse model. Kit^V558Δ/+ mice have an average lifespan of 8 months due to progressive bowel obstruction (Figure 4). Tumors from Kit^V558Δ/+ mice express canonical markers of GIST including the tyrosine kinase KIT and the transmembrane channel DOG1 (Figure 5), as well as the transcription factor ETV1 (not shown). Tumors can be studied for changes in the KIT signaling pathway, such as the downstream markers ERK and AKT (Figure 6), or immune microenvironment which closely mimics human GIST^8,11,12,13. MRI⁸ or CT (Figure 7) can also be used to track tumor volume as an accurate measurement of tumor response. An untreated Kit^V558Δ/+ mouse was given 200 µL of gastrograffin orally 1 h prior to imaging. CT imaging was completed on a vivaCT 80 platform. 3D reconstruction was performed with Fiji software, which can also measure tumor volume.

Figure 1: Comparison of tumor weights in male and female Kit^V558Δ/+ mice. Tumors from 9-week-old untreated Kit^V558Δ/+ male and female mice were isolated and weighed (n = 15 mice/group). Data represents mean ± standard error of mean (SEM); p values were calculated using a student's t test; * = P < 0.05. Please click here to view a larger version of this figure.

Figure 2: Effect of imatinib on tumors from Kit^V558Δ/+ mice. Kit^V558Δ/+ mice were treated with vehicle or 600 mg/L imatinib in drinking water for 1 or 4 weeks. Tumors were isolated and weighed (n = 4-5 mice/group). Data represents mean ± SEM; p values were calculated using a one-way ANOVA comparison with a Bonferroni post-test for comparison of individual groups; * = P < 0.05. Please click here to view a larger version of this figure.

Figure 3: Kit^V558Δ/+ tumor with cap. Representative photo of a tumor from a Kit^V558Δ/+ mouse with a cecal cap containing serous fluid. Please click here to view a larger version of this figure.

Figure 4: Lifespan of Kit^V558Δ/+ mice. Survival of untreated Kit^V558Δ/+ mice was tracked for > 400 days (n = 43 mice). Please click here to view a larger version of this figure.

Figure 5: Immunohistochemical analysis. Representative histology of a Kit^V558Δ/+ tumor where scale bar is 40 µm. Abbreviations: H&E = hematoxylin and eosin staining; Kit = a canonical mark of GIST and receptor tyrosine kinase; Dog1 = a canonical marker of GIST with role in anion transport. Please click here to view a larger version of this figure.

Figure 6: Molecular signaling analysis. Kit^V558Δ/+ mice were treated with vehicle or 600 mg/L imatinib in drinking water for 1 week. Mice were given a single i.p. injection of vehicle or 45 mg/kg imatinib 6 h prior to analysis. Protein lysates from Kit^V558Δ/+ tumors were examined by western blot (n = 2 mice/group). Abbreviations: P Kit = phosphorylated Kit receptor tyrosine kinase; T kit = total Kit receptor tyrosine kinase; P ERK = phosphorylated mitogen activated protein kinase; T ERK = total mitogen activated protein kinase; P AKT = phosphorylated serine-threonine protein kinase; T AKT = total serine-threonine protein kinase. Please click here to view a larger version of this figure.

Figure 7: CT imaging analysis. 3D CT reconstruction of an untreated Kit^V558Δ/+ mouse demonstrating a tumor (arrow) in the pelvis. Please click here to view a larger version of this figure.

The Kit^V558Δ/+ mouse model is a powerful research tool in the molecular and immunologic analysis of GIST. Although the breeding strategy requires a single cross, using Kit^V558Δ/+ mouse cohorts in experiments analyzing tumor response requires extensive breeding. Mice should be age- and sex-matched to ensure similar tumor weights, and 10% of mice die before 8 weeks of age when tumors are established. Less extensive breeding strategies are possible if using advanced imaging techniques such as CT or MRI to track tumor volume within individual mice. Nonetheless, Kit^V558Δ/+ mice have been successfully crossed to other knock-out or inducible mouse models, revealing important immune and molecular mechanisms¹².

Tumor cells from Kit^V558Δ/+ mice are easily isolated by column sorting for KIT (CD117) or by flow cytometry¹⁸. Isolated tumor cells from Kit^V558Δ/+ mice can be grown in vitro¹². At early passages, isolated tumor cells from Kit^V558Δ/+ mice retain KIT expression and can be used for in vitro studies. However, after several passages, these cell lines lose KIT expression, limiting their applicability. Like most mouse models, Kit^V558Δ/+ tumors do not metastasize, which limits the study of extraintestinal GIST. Similarly, tumors from Kit^V558Δ/+ mice only occur in the cecum, and cells isolated from Kit^V558Δ/+ tumors do not grow when isolated and injected into the liver or spleen, which limits the evaluation of the tumor microenvironment from different sites of disease. Also, the Kit^V558Δ/+ mutation does not have any known impact on the hematological development of cells in this mouse model.

The immune microenvironment in tumors from Kit^V558Δ/+ mice contains mostly macrophages followed by T cells, which closely mimics human GIST¹¹. However, it is imperative that the cecal cap be removed completely from the tumor prior to assessing tumor weight or immune infiltrate, as distinct immune populations exist in the tumor versus the cecum. In our experience, tumor weights are comparable even among those with caps, and there is no significant association between tyrosine kinase inhibitor therapy and the development of a cecal cap. Furthermore, we have found no significant difference in the tumor microenvironment comparing tumors with and without cecal caps.

Many genetically engineered mouse models have been developed for use in cancer research, especially since the advent of CRISPR gene editing. While numerous immunocompetent models rely on cre-loxP mediated activation of oncogenes or inactivation of tumor suppressor genes, the Kit^V558Δ/+ tumors are driven by their endogenous promoter. Accordingly, findings in the Kit^V558Δ/+ mouse model are highly translatable to human disease, particularly in the evaluation of KIT signaling¹⁹. Looking ahead, the Kit^V558Δ/+ mouse should continue to serve as a valuable model as novel treatments strategies including checkpoint blockade and CAR T therapy continue to be explored for the treatment of soft tissue tumors including GIST.