Gene expression differences in Mertk ^-/- RPE revealed by genome-wide transcriptional analyses in the presence of 129P2 versus B6 alleles on chromosome 2

A segment of chromosome 2 in the Mertk ^-/-V1 mice was previously reported to be derived from the 129P2 background (Vollrath et al., 2015). Therefore, we performed short tandem repeat (STR) analysis in the chromosome 2 region surrounding the Mertk locus in B6 WT, Mertk ^-/-V1, Mertk ^-/-V2, and Mertk ^-/-V3 mice. Genomic DNA isolated from each of these mouse lines was subjected to PCR amplification of 24 microsatellite sites across chromosome 2 (Figure 3—figure supplement 1). We found that Mertk ^-/-V1 mice harbored a 15.08 cM region between D2Mit206 and D2Mit168 that is of 129P2 origin (Figure 3A). As expected, chromosome 2 was entirely B6-derived in both Mertk ^-/-V2 and Mertk ^-/-V3 mice (Figure 3A).

Open in a separate window

Figure 3.

Significant gene expression differences in retinal pigment epithelia (RPE) of Mertk ^-/-V1 versus Mertk ^-/-V2 and Mertk ^-/-V3 mice.

(A) Haplotype map comparing 20 microsatellite markers across chromosome 2 in C57BL/6, Mertk ^-/-V1, Mertk ^-/-V2, and Mertk ^-/-V3 mice. Black rectangles indicate homozygosity for C57BL/6 (B6) alleles, and gray rectangles indicate homozygosity for 129P2/OlaHsd (129P2) alleles. (B) Distribution of differentially expressed genes (DEGs) in RPE across chromosomes (n = 3–6 samples/genotype). *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001, hypergeometric test. (C) Schematic showing genes neighboring Mertk that are significantly upregulated or downregulated in the RPE of Mertk ^-/-V1 mice. (D) qPCR quantification of indicated chromosome 2 genes in C57BL/6, Mertk ^-/-V1, Mertk ^-/-V2, and Mertk ^-/-V3 RPEs (mean ± SEM, n = 4–5 samples/genotype). *p<0.05, **p<0.01, ***p<0.001, one-way ANOVA Dunnett’s test. Source files for the distribution of DEGs in RPEs across chromosomes (B) and qPCR quantification of various chromosome 2 genes (D) are available in Figure 3—source data 1. Supporting data for (A) is available in Figure 3—figure supplement 1.

Figure 3—source data 1.

Independent datasets for qPCR quantifications shown in Figure 3D.

Click here to view.^{(22K, xlsx)}

Figure 3—figure supplement 1.

Open in a separate window

A 129P2-specific genomic interval is present in Mertk ^-/-V1 but not in Mertk ^-/-V2 and Mertk ^-/-V3 mice.

Representative PCR amplification products corresponding to 24 different microsatellite markers on chromosome 2 using genomic DNA extracted from C57BL/6, Mertk ^-/-V1, Mertk ^-/-V2, and Mertk ^-/-V3 mice. The presence of a 129P2/OlaHsd (129P2)-derived segment of DNA on chromosome 2 in Mertk ^{-/- V1} mice was confirmed, at each microsatellite location shown, in at least three animals. Source files for the gel images shown are available in .

Figure 3—figure supplement 1—source data 1.

Unmodified images for PCR genotyping results shown in Figure 3—figure supplement 1.

Click here to view.^{(18M, zip)}

To more broadly understand the genome-wide transcriptional differences between Mertk ^-/-V1, Mertk ^-/-V2, and Mertk ^-/-V3 mice stemming from the chromosomal differences, we performed RNA sequencing (RNAseq) experiments on RPE from these three mouse lines and control B6 WT RPE at P25. We identified a number of genes that were differentially expressed in Mertk ^-/-V1, Mertk ^-/-V2, and Mertk ^-/-V3 RPE in cis and in trans compared to B6 WT RPE (Figure 3B and C). Significant changes in the transcripts of Mertk-neighboring genes, in Mertk ^-/-V1, but not Mertk ^-/-V2 and Mertk ^-/-V3, RPE relative to B6 WT RPE were confirmed by qPCR (Figure 3D).

Tyro3 is epistatic with Mertk for the retinal degeneration trait

It was previously reported that TYRO3 levels were significantly lower in Tyro3 ^129/129 RPE compared to Tyro3 ^B6/B6 or Tyro3 ^129/B6 RPE (Vollrath et al., 2015). Consistent with this report, we found Tyro3 mRNA level to be significantly downregulated in Mertk ^-/-V1, but not in Mertk ^-/-V2 and Mertk ^-/-V3, RPE (Figure 4A). Moreover, we detected significantly lower levels of TYRO3 in Mertk ^-/-V1 RPE by Western blot (Figure 4B). By contrast, RPE cells from Mertk ^-/-V2 and Mertk ^-/-V3 mice had levels of TYRO3 that were comparable to B6 WT mice (Figure 4B). Since it was concluded that even the hypomorphic expression of Tyro3 ^B6/129 can suppress the phenotypes of Mertk loss of function (Vollrath et al., 2015), we tested whether the simultaneous genetic ablation of Mertk and Tyro3 can phenocopy Mertk ^-/-V1 mice. We engineered Mertk ^-/-V2 Tyro3 ^-/-V2 mice by targeting Tyro3 in Mertk ^-/-V2 mice using CRISPR/CAS9 (Figure 4C, Figure 1—figure supplement 1D). Immunoblotting experiments confirmed that TYRO3 was undetectable in the RPE of Mertk ^-/-V2 Tyro3 ^-/-V2 mice when compared to TYRO3 expression in B6 WT RPE (Figure 4D and E). Indeed, these mice recapitulated the severe retinal degeneration observed in Mertk ^-/-V1 underscoring the function of Tyro3 ^B6 as a suppressor allele in retinal degeneration induced by targeting Mertk. When eye sections from 6-month-old Mertk ^-/-V2 Tyro3 ^-/-V2, Mertk ^-/-V1 and B6 WT controls were stained with hematoxylin-eosin, we found that ONL thickness in Mertk ^-/-V2 Tyro3 ^-/-V2 mice was significantly reduced compared to B6 WT controls, commensurate with Mertk ^-/-V1 mice (Figure 4F and G). Similar to Mertk ^-/-V1 mice, Mertk ^-/-V2 Tyro3 ^{-/- V2} mice had only approximately one row of nuclei in the ONL across the entire dorsal–ventral axis of the retina (Figure 4F and G). Consequently, Mertk ^-/-V2 Tyro3 ^-/-V2 mice did not display light-evoked responses in scotopic ERGs at any luminance tested (Figure 4H and I). These results show that Mertk ^-/-V2 Tyro3 ^-/-V2 mice phenocopy the widespread morphological and functional retinal deficits characteristic of Mertk ^-/-V1 mice.

Open in a separate window

Figure 4.

Tyro3 is epistatic with Mertk for the retinal degeneration trait.

(A) qPCR quantification of Tyro3 in C57BL/6, Mertk ^-/-V1, Mertk ^-/-V2, and Mertk ^-/-V3 retinal pigment epithelia (RPE) (mean ± SEM, n = 4–5 samples/genotype). **p<0.01, one-way ANOVA Dunnett’s test. (B) Representative and independent measurements of TYRO3 amounts in RPE. Western blot (WB) from C57BL/6, Mertk ^-/-V1, Mertk ^-/-V2, Mertk ^-/-V3, and Tyro3 ^-/-V1 mice RPE (mean ± SEM of n = 5 mice/ genotype). *p<0.05, **p<0.01, ***p<0.001, and ****p<0.0001, one-way ANOVA Dunnett’s test. (C) Schematic showing targeting of Tyro3 exons 7–18 with CRISPR/Cas9 in Mertk ^-/-V2 ES cells to generate the Mertk ^{-/- V2} Tyro3 ^{-/- V2} mouse line. Image not drawn to scale. (D, E) Representative and independent measurements of TYRO3 amounts in RPE. WB from C57BL/6, Mertk ^-/-V2 Tyro3 ^-/-V2, and Tyro3 ^-/-V1 mice RPE (mean ± SEM of n = 4 mice/ genotype). ****p<0.0001, one-way ANOVA Dunnett’s test. (F) Representative hematoxylin-eosin-stained transverse sections of the retina. Boxed section is shown as inset and indicates the areas quantified in (G). Scale bars = 200 mm (left panels) and 10 mm (insets). (G) Quantification of outer nuclear layer (ONL) thickness in the area indicated in (F) (mean ± SEM of 10 measurements/mouse, n = 3–6 mice/genotype). ****p<0.0001, one-way ANOVA Dunnett’s test. (H) Representative scotopic electroretinogram traces are shown at the highest luminance tested. (I) Quantification of a-wave amplitude and b-wave amplitude at highest luminance tested (mean ± SEM of n = 3–6 mice/genotype). *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001, one-way ANOVA Dunnett’s test. a-wave amplitude at increasing luminances. **p<0.01, ****p<0.0001, two-way ANOVA. Morphological and functional changes in the eye were assessed in 6-month-old C57BL/6, Mertk ^-/-V1, and Mertk ^-/-V2 Tyro3 ^-/-V2 mice. Source files for (A) qPCR quantification of Tyro3, (B, E) quantification of TYRO3 levels, (G) ONL thickness, (I) a-wave amplitude, b-wave amplitude, and a-wave amplitude at increasing luminances are available in Figure 4—source data 1. Supporting data for (C) is available in Figure 1—figure supplement 1D. GCL, ganglion cell layer; IPL, inner plexiform layer; INL, inner nuclear layer; OPL,-outer plexiform layer.

Figure 4—source data 1.

Independent datasets and unmodified images for results shown in Figure 4.

Click here to view.^{(1.5M, zip)}

Mertk targeting in 129P2 ES cells results in a robust anti-tumor response against immune checkpoint inhibitor (ICI)-refractory YUMM1.7 melanoma and GL261 brain tumor, but this effect is not phenocopied in Mertk ^-/-V2 or Mertk ^-/-V3 mice

An important, newly discovered role of MERTK is as an innate immune checkpoint in cancer (Cook et al., 2013; Davra et al., 2021; Huelse et al., 2020; Lee-Sherick et al., 2020; Lin et al., 2022; Lindsay et al., 2021; Sekar et al., 2022; Su et al., 2020; Wu et al., 2018; Zhou et al., 2020). Mertk ^-/-V1 mice were used to demonstrate improved anti-tumor immune response against MMTV-PyVmT, B16:F10, and MC38 tumor models (Cook et al., 2013). Based on the use of Mertk ^-/-V1 mice, it was also surmised that inhibition of macrophage efferocytosis due to MERTK loss of function results in decreased tumor growth and increased tumor-free survival (TFS) in E0771 murine breast cancer model (Davra et al., 2021). Similarly, Mertk ^-/-V1 mice were used in a study by Lindsay et al. to demonstrate improved anti-tumor T cell motility in a B78ChOva tumor model (Lindsay et al., 2021). We compared the anti-tumor response of Mertk ^-/-V1, Mertk ^-/-V2, and Mertk ^-/-V3 mice in a model of ICI-refractory melanoma (YUMM1.7), as well as in an orthotopic brain tumor mouse model (GL261). YUMM1.7 tumor cells were implanted subcutaneously in Mertk ^-/-V1, Mertk ^-/-V2, and Mertk ^-/-V3 mice or in B6 WT (Figure 5A). TFS, overall survival (OS), and rate of tumor growth were monitored (Figure 5A and B). Consistent with previous reports of MERTK blockade enhancing anti-tumor response (Cook et al., 2013; Davra et al., 2021; Huelse et al., 2020; Lee-Sherick et al., 2020; Lin et al., 2022; Lindsay et al., 2021; Sekar et al., 2022; Su et al., 2020; Wu et al., 2018; Zhou et al., 2020), 93.75% of Mertk ^-/-V1 mice remained tumor free compared to 0% of B6 WT control mice (Figure 5B). Surprisingly, neither of the Mertk knockout mice generated using B6 ES cells, Mertk ^-/-V2 and Mertk ^-/-V3, phenocopied the tumor resistance observed in Mertk ^-/-V1 mice. Also, 100% of Mertk ^-/-V2 and Mertk ^-/-V3 mice succumbed to tumor growth at a rate comparable to their B6 WT counterparts (Figure 5B). To investigate these findings in an independent tumor model, luciferase-expressing GL261 brain tumor cells were orthotopically injected in B6 and Mertk ^-/-V1 mice. Bioluminescence imaging was performed to determine tumor volume and mice were monitored for OS (Figure 5—figure supplement 1). No differences were observed in the rate of GL261 growth or time to end point between B6 WT or Mertk ^-/-V1 mice (Figure 5—figure supplement Figure 5—figure supplement 1B and C). We next attempted to treat tumor-bearing mice with a dendritic cell vaccine (DC-Vax). Tumor cell lysates obtained by freeze-thawing GL261 were fed to B6 WT bone marrow-derived DCs. Following co-incubation, DCs were activated with LPS treatment and intraperitoneally injected into tumor-bearing B6 WT or Mertk ^-/-V1 mice on days 14 and 21 after tumor implantation (Figure 5C). Of note, tumor sizes were comparable between B6 WT or Mertk ^-/-V1 mice prior to receiving the DC-Vax (Figure 5E, top panel). Remarkably, the administration of a DC-Vax in Mertk ^-/-V1 mice resulted in a significant reduction in tumor size, whereas DC-Vax-treated B6 WT mice succumbed to their tumor burden (Figure 5D and E). Similar to the studies on YUMM1.7 melanoma cells, when GL261 cells were implanted in Mertk ^-/-V2 and Mertk ^{-/- V3} mice, and these mice were subsequently treated with DC-Vax, 100% of these mice failed to display the anti-GL261 response of Mertk ^-/-V1 mice (Figure 5D and E). Taken together, our experiments revealed an astounding anti-tumor resistance of Mertk ^-/-V1 mice against both YUMM1.7 and GL261. However, this phenotype of Mertk ^-/-V1 mice was again not phenocopied by Mertk ^-/-V2 and Mertk ^-/-V3 mice.

Open in a separate window

Figure 5.

Anti-tumor response of Mertk ^-/-V1 mice is not phenocopied by Mertk ^-/-V2 and Mertk ^-/-V3 mice.

(A) Schematic showing subcutaneous injection of 100,000 Braf ^V600E Pten ^-/- YUMM1.7 mouse melanoma cells into mice of different genotypes. (B) Tumor-free survival (TFS), overall survival (OS), and tumor volume in C57BL/6 (n = 17), Mertk ^-/-V1 (n = 16), Mertk ^-/-V2 (n = 9), and Mertk ^-/-V3 (n = 7) mice implanted with YUMM1.7 cells. ****p<0.0001, Log-rank Mantel–Cox test or two-way ANOVA, Dunnett’s multiple-comparison test. (C) Schematic showing intracranial injection of 10,000 GL261-Luc glioma cells and intraperitoneal dendritic cell vaccination of mice at 14 and 21 days post-tumor implantation. (D) OS in C57BL/6 (n = 16), Mertk ^-/-V1 (n = 10), Mertk ^-/-V2 (n = 8), and Mertk ^{-/- V3} (n = 8) mice implanted with GL261 tumors. ****p<0.0001, Log-rank Mantel–Cox test. (E) Representative IVIS images of intracranial tumors at D14 and D28 post-implantation. Source files for TFS (B) and OS (B, D) plots shown are available in Figure 5—source data 1. Supporting data relating to (C–E) is available in Figure 5—figure supplement 1.

Figure 5—source data 1.

Tumor free survival dataset following implantation with YUMM1.7 melanoma cells shown in Figure 5.

Click here to view.^{(14K, xlsx)}

Figure 5—source data 2.

Overall survival dataset following implantation with GL261 glioblastoma cells shown in Figure 5.

Click here to view.^{(9.9K, xlsx)}

Figure 5—figure supplement 1.

Open in a separate window

Unvaccinated Mertk ^{-/- V1} mice are not protected against GL261 tumors.

(A) Schematic of intracranial injection of 10,000 GL261-Luc glioma cells in mice. (B) Overall survival (OS) in C57BL/6 (n = 9) and Mertk ^-/-V1 (n = 11) mice. n.s., Log-rank Mantel–Cox test. (C) Representative IVIS images of intracranial tumors in mice at D14 and D28 post-implantation. Source file for the OS plot (B) shown is available in Figure 5—source data 2.

Neither deficient efferocytosis in macrophages nor loss of Tyro3 ^B6/B6 can universally account for the anti-tumor immunity in Mertk ^-/-V1 mice

The prevailing dogma is that deficient phagocytosis of tumor cells by macrophages in the absence of Mertk function represents the sine qua non of the remarkable anti-tumor immunity. It was proposed that the absence of MERTK in tumor-associated macrophages prevents the phagocytosis and disposal of dead tumor cells, thereby increasing the availability of tumor antigen to DCs and triggering improved anti-tumor T cell immunity (Davra et al., 2021). Similarly, it was speculated that deficient MERTK-dependent phagocytosis by tumor-associated macrophages leads to secondary necrosis of tumor cells and an increased pro-inflammatory microenvironment more conducive to anti-tumor immunity (Stanford et al., 2014). It was also suggested that blockade of MERTK-dependent phagocytosis in tumor-associated macrophages activated the STING pathway (Davra et al., 2021; Stanford et al., 2014; Zhou et al., 2020). Of note, these scenarios are not mutually exclusive and might in fact cooperate. We performed similar RNAseq analyses in BMDMs isolated from Mertk ^-/-V1, Mertk ^-/-V2 and Mertk ^-/-V3 and B6 WT mice. Our experiments identified a number of changes in Mertk ^-/-V1, Mertk ^-/-V2, and Mertk ^-/-V3 BMDMs compared to B6 WT controls. Furthermore, Mertk ^-/-V1 mice had a significant number of upregulated and downregulated genes that were not recapitulated in Mertk ^-/-V2 and Mertk ^-/-V3 mice (Figure 6A–C). Although the lack of anti-tumor resistance in Mertk ^-/-V2 and Mertk ^-/-V3 mice already pointed to a MERTK-agnostic basis for tumor clearance, we hypothesized that some of the coincidental changes in BMDMs in these mice might compensate for the loss of MERTK by providing redundancy in phagocytosis. Thus, we expected no reduction in efferocytosis in BMDMs derived from Mertk ^-/-V2 and Mertk ^-/-V3 mice, unlike BMDMs from Mertk ^-/-V1 mice. We generated CD11b⁺ F4/80⁺ BMDMs from B6 WT, Mertk ^-/-V1, Mertk ^-/-V2, and Mertk ^-/-V3 mice and tested them in an ex vivo phagocytosis assay. Flow cytometry-based analysis of BMDMs derived from Mertk ^-/-V2 and Mertk ^-/-V3 mice confirmed MERTK ablation in these cells compared to B6 controls (Figure 6—figure supplement Figure 6—figure supplement 1A and B). BMDMs were cultured with pHrodo-labeled apoptotic thymocytes in the presence of serum and their uptake was assayed by flow cytometry after 1 hr. As expected, Mertk ^-/-V1 BMDMs were ~50% less phagocytic than B6 WT BMDMs. BMDMs from Mertk ^-/-V2 and Mertk ^-/-V3 mice were similarly less phagocytic than those from B6 WT mice (Figure 6D and E). These experiments demonstrate that BMDMs derived from Mertk knockout mice of 129P2 or B6 origin were equally deficient in the efferocytosis of apoptotic thymocytes.

Open in a separate window

Figure 6.

Anti-tumor response of Mertk ^-/-V1 mice is neither the result of deficient efferocytosis by macrophages nor hypomorphic TYRO3.

(A) Distribution of differentially expressed genes (DEGs) in bone marrow-derived macrophages (BMDMs) across chromosomes. (n = 3 samples/genotype). *p<0.05, ****p<0.0001, hypergeometric test. (B) Schematic indicating chromosome 2 genes that are upregulated or downregulated in Mertk ^-/-V1 BMDMs. (C) qPCR quantification of the indicated chromosome 2 genes in C57BL/6, Mertk ^-/-V1, Mertk ^-/-V2, and Mertk ^-/-V3 BMDMs (mean ± SEM, n = 5 samples/genotype). *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001, one-way ANOVA Dunnett’s test. (D) BMDMs from C57BL/6, Mertk ^-/-V1, Mertk ^-/-V2, and Mertk ^-/-V3 mice were co-cultured with apoptotic thymocytes for 1 hr. Representative plots show uptake of pHrodo-labeled apoptotic thymocytes by CD11b⁺F4/80⁺ BMDMs by flow cytometry. (E) Quantification of (D) (mean ± SEM, n = 3–6 mice/group). **p<0.01, Mann–Whitney or Student’s t-test. (F) Tumor-free survival (TFS), overall survival (OS), and tumor volume in C57BL/6 (n = 7), Mertk ^-/-V1 (n = 7), and Mertk ^-/-V2 Tyro3 ^-/-V2 (n = 7) mice implanted with YUMM1.7 tumors. n.s., Log-rank Mantel–Cox test. (G) OS in C57BL/6 (n = 5) and Mertk ^-/-V2 Tyro3 ^-/-V2 (n = 6) mice implanted with GL261 tumors. n.s., Log-rank Mantel–Cox test. (H) Representative IVIS images of intracranial tumors in C57BL/6 and Mertk ^-/-V2 Tyro3 ^-/-V2 mice at D14 and D28 post-implantation. Source files for the distribution of DEGs in BMDMs across chromosomes (A), qPCR quantification of various chromosome 2 genes (C), quantification of percentage of pHrodo⁺ BMDMs (E), TFS (F), and OS (F, G) plots are available in Figure 6—source data 1. Supporting data for (D, E) is available in Figure 6—figure supplement 1.

Figure 6—source data 1.

Independent datasets for qPCR quantifications and survival dataset following implantation with tumor models shown in Figure 6.

Click here to view.^{(25K, xlsx)}

Figure 6—source data 2.

Quantification of MERTK levels in bone-marrow-derived macrophages shown in Figure 6.

Click here to view.^{(9.4K, xlsx)}

Figure 6—figure supplement 1.

Open in a separate window

Significantly diminished MERTK expression in bone marrow-derived macrophages (BMDMs) isolated from Mertk knockout mice.

(A, B) Representative flow cytometry and independent quantification of MERTK levels in BMDMs from C57BL/6, Mertk ^-/-V1, Mertk ^-/-V2, and Mertk ^-/-V3 mice (mean ± SEM of n = 4–5 mice/genotype). ****p<0.0001, one-way ANOVA Dunnett’s test; **p<0.01, Mann–Whitney test. Source files for quantification of MERTK levels in BMDMs are available in Figure 6—source data 2.

It is important to note that TYRO3 can not only compensate for MERTK in phagocytosis (Vollrath et al., 2015); it is also a negative regulator of the immune response (Chan et al., 2016). However, Tyro3 was not a differentially expressed gene (DEG) in BMDMs because it is not expressed in these cells (Figure 6A–C). To entirely rule out Tyro3 ^B6 function as a suppressor in abolishing the anti-tumor effects of Mertk ablation in Mertk ^-/-V2 and Mertk ^-/-V3 mice, we implanted YUMM1.7 and GL261 tumors in Mertk ^-/-V2 Tyro3 ^-/-V2 mice. However, we observed that Mertk ^-/-V2 Tyro3 ^-/-V2 mice entirely failed to phenocopy the anti-tumor resistance observed in Mertk ^-/-V1 mice (Figure 6F–H). Specifically, 100% of Mertk ^-/-V2 Tyro3 ^-/-V2 mice failed to show improved survival in comparison to B6 WT mice when implanted with YUMM1.7 (Figure 6F). No differences were observed in TFS, OS, or in tumor growth (Figure 6F). Similarly, in the GL261 model, no differences were observed in OS or in tumor volume in Mertk ^-/-V2 Tyro3 ^-/-V2 mice following the DC vaccination strategy that conferred significant anti-tumor resistance to Mertk ^-/-V1 mice (Figure 6G and H).

Animals

Animals were bred and maintained under a strict 12 hr light cycle and fed with standard chow diet in a specific pathogen-free facility at Yale University. All experiments involving animals were performed in accordance with regulatory guidelines and standards set by the Institutional Animal Care and Use Committee of Yale University. All C57BL/6 mice were purchased from Jackson Laboratories and subsequently bred and housed at Yale University. The widely used Mertk ^-/-V1 mice have been described previously (Camenisch et al., 1999; Lu et al., 1999). Mertk ^-/-V1 mice were crossed onto the C57BL/6J background (Jackson Laboratories strain# 000664) for at least 10 generations. As shown in Figure 1, Mertk ^-/-V1 mice have deletion of exon 17 that corresponds to the kinase domain of Mertk and the neomycin cassette is still present in the Mertk locus. Of note, exon 17 was referred to as exon 18 in the original description of the mouse (Lu et al., 1999).

To generate an independent line of mice that has Mertk deleted globally (referred to as Mertk ^-/-V2 mice), we bred Mertk ^f/f mice with commercially available Rosa26^ERT2Cre⁺ (Jackson Laboratories strain# 008463) mice. A description detailing the generation of the Mertk ^f/f mice can be found in Fourgeaud et al., 2016 (Fourgeaud et al., 2016). Mertk ^f/f mice originated from embryos of C57BL/6NJ background. Germline inactivation of the Mertk ^f/f allele in Mertk ^f/f Rosa26^ERT2Cre⁺ mice was achieved by intraperitoneally injecting 3 mg of 4-hydroxytamoxifen (Sigma-Aldrich H7904) for five consecutive days. Two weeks post-tamoxifen injection, adult males and females were set up as breeder pairs. Litters from this cross were then screened for excision of the Mertk ^f/f allele. Once identified, excision positive mice were bred with C57BL/6J mice to eliminate the Cre recombinase. Finally, mice were genotyped to set apart Mertk ^-/-V2 founder mice that had excised exon 18, encoding for the kinase domain of Mertk, on both alleles. Genome-wide single-nucleotide polymorphism (SNP) analysis of our established Mertk ^-/-V2 mice indicated that ~84.37% of their genome was C57BL/6J-derived while ~15.62% was C57BL/6NJ-derived. An independent line targeting deletion of Mertk, designated as Mertk ^-/-V3 mice, was generated at Cyagen Biosciences Inc (Santa Clara, CA), by CRISPR/Cas9-mediated genome engineering. As shown in Figure 1, Mertk ^-/-V3 mice have exons 3 and 4 targeted; transcribed mRNA from targeted allele with frameshift mutation undergo nonsense-mediated decay. Single-guide RNAs (sgRNAs) were injected into fertilized C57BL/6NJ eggs, and founder animals were identified by PCR followed by DNA sequencing analysis. Genome-wide SNP analysis revealed that ~55.22% of their genome was C57BL/6J-derived and ~44.73% was C57BL/6NJ-derived.

CRISPR/Cas9 technique was used to make Mertk ^-/-V2 Tyro3 ^-/-V2 mice, as described previously (Henao-Mejia et al., 2016), at the Yale Genome Editing Center. In brief, T7-sgRNA templates were prepared by PCR, incorporating the guide sequences from the desired target regions in the mouse Tyro3 gene (NCBI Gene ID: 22174), with a 5′ guide sequence of CTACACCTACAGAGAACAAG (sense orientation, cutting the gene at 8387/8 bp within intron 6) and a 3’ guide sequence of CCCAAGTGTCAGAATCCCAG (sense orientation, cutting the gene at 17,800/1 bp within intron 18), thus resulting in a deletion of 9413 bp. The T7-sgRNA PCR templates were then used for in vitro transcription and purification with the MEGAshortscript T7 Transcription Kit and MEGAclear Transcription Clean-Up Kit, respectively (both from Thermo Fisher Scientific AM1354, AM1908). Cas9 mRNA (CleanCap, 5-methoxyuridine-modified) was purchased from TriLink Biotechnologies. Subsequently, cytoplasmic microinjections of sgRNAs and Cas9 mRNA into single-cell embryos obtained from Mertk ^-/-V2 donors were performed. Founder mice with heterozygous deletion of the Tyro3 allele were identified with genotyping by PCR. Next, Mertk ^-/-V2 Tyro3 ^-/-V2 line was established by heterozygote-to-heterozygote breeding. Consistent with previous reports (Chen et al., 2009; Lu et al., 1999), male double knockout mice lacking both Mertk and Tyro3 have significantly smaller testicles and reduced fertility. Tyro3 ^-/-V1 mice have been described previously (Lu et al., 1999).

PCR amplification

PCR reactions for sequencing were performed using the primers listed in Appendix 1—table 1. PCR amplifications were carried out with TopTaq Master Mix Kit (QIAGEN). PCR reactions of 25 μl were performed with 2 μl genomic DNA, 0.2 μM primer pair, 2.5 μl CoralLoad Concentrate 10×, 12.5 μl TopTaq Master Mix, 2× (contains TopTaq DNA polymerase, TopTaq PCR Buffer with 3 mM MgCl₂ and 400 μM each dNTP). Thermal cycling conditions were based on Touchdown PCR method described by Korbie and Mattick, 2008. PCR products were examined by gel electrophoresis.

STR analysis

Genomic DNA was isolated from liver biopsies using the DNeasy blood and tissue kit (QIAGEN 69504) according to the manufacturer’s protocol. PCR reactions to amplify 24 different microsatellite regions on chromosome 2 were performed using the primers listed in Appendix 1—table 1. Thermal cycling conditions were optimized based on Touchdown PCR method described by Korbie and Mattick, 2008. Following PCR amplification, STR markers were scored by resolving PCR products on 4% agarose gels.

Western blot

Protein lysates from adult mice RPE were obtained using a previously validated protocol (Wei et al., 2016). Briefly, the neural retina was removed and posterior eyecups were incubated on ice in 200 μl of RIPA buffer with protease inhibitor cocktail, EDTA-free (Roche 11697498001) for up to 1 hr. Posterior eyecups were removed and the dislodged RPE cells were sonicated for 20 s on ice. After 10 min at 14,000 rpm in a refrigerated centrifuge, supernatants were transferred to new tubes and protein content was quantified with Pierce BCA assay (Thermo Fisher Scientific 23227) as per the manufacturer’s instructions. Concomitantly, spleen, brain, testes, and peritoneal exudate were collected from adult mice. Samples were kept in NP-40 buffer containing a cocktail of protease inhibitors (Roche 11697498001). Tissues were mechanically disrupted and left rotating at 4°C for 2 hr to ensure complete homogenization. Subsequently, all samples were centrifuged for 20 min at 12,000 rpm and the supernatants were collected for protein quantification as described above.

For immunoblots, equal amounts of total protein in Laemmli buffer were subjected to electrophoresis on precast polyacrylamide gels (Bio-Rad 4568025, 4568125) and transferred to PVDF membranes (Bio-Rad 1620177). Membranes were blocked and probed overnight with corresponding primary antibodies (MERTK: Abcam ab95925 1/1000, TYRO3: Cell Signaling Technology 5585S 1/1000, β-actin: Cell Signaling Technology 8457S and 3700S). Secondary antibodies conjugated to near-infrared fluorophores (LI-COR Biosciences 926-32213, 926-68072) were detected using Odyssey Classic Imaging System (LI-COR Biosciences) and quantified with Image Studio Lite Software (LI-COR Biosciences).

Generation of BMDMs

BMDMs from age-matched, adult C57BL/6 Mertk ^-/-V1, Mertk ^-/-V2, and Mertk ^-/-V3 mice were differentiated from bone marrow precursors. Briefly, bone marrow cells were isolated and propagated for 7 days in 30% L929-conditioned RPMI (Gibco 11875101) containing 20% FCS (Sigma-Aldrich 18D078) and 1% Penicillin-Streptomycin (Gibco 10378016). At day 7, BMDMs were lifted for downstream assays.

Phagocytosis assays

Thymocytes were isolated from 3- to 6-week-old C57BL/6 mice. Thymocytes were incubated for 4 hr with 1 μg/ml of dexamethasone (Sigma-Aldrich D4902) to induce apoptosis. In parallel, BMDMs from the indicated mice were collected and replated to adhere for 3 hr at 37°C. Apoptotic thymocytes, pre-labeled with 0.1 mg/ml of pHrodo-SE (Thermo Fisher P36600), were subsequently added to BMDMs at 6:1 ratio. Cells were co-cultured for 1 hr at 37°C. Afterward, the apoptotic thymocyte-containing media were removed and BMDMs were washed five times with 1× PBS. Adherent BMDMs were then treated with Accutase (Sigma-Aldrich A6964) for 10 min at 37°C. BMDMs were then gently scraped off for collection and stained for analysis with flow cytometry.

Flow cytometry staining and acquisition

Single-cell suspensions of BMDMs were stained in PBS with fixable viability dye (Thermo Fisher Scientific 65-0865-14) for 10 min. Cells were then incubated in 2% FCS/PBS solution containing anti-mouse CD16/32 antibody (BioLegend clone 93) for 15 min and subsequently stained with a combination of fluorophore-conjugated primary antibodies against mouse CD11b (BioLegend clone M1/70), F4/80 (BioLegend clone BM8), and MERTK (Invitrogen clone DS5MMER) at 4°C for 25 min. After staining, cells were washed and data was immediately acquired with BD LSRII flow cytometer using BD FACSDiva software (BD Biosciences). Finally, raw data were analyzed using FlowJo software (Tree Star Inc).

Histological analysis

After mice were sacrificed by carbon dioxide inhalation, eyecups were immediately collected and incubated overnight in eye fixative (ServiceBio). Hematoxylin and eosin staining was performed by iHisto Inc Samples were processed, embedded in paraffin, and sectioned at 4 μm. Paraffin sections were then deparaffinized and hydrated using the following steps: xylene, two rounds,15 min each; 100% ethanol, two rounds, 5 min each; 75% ethanol, one round, 5 min; and 1× PBS, three rounds, 5 min each at room temperature. After deparaffinization, 4‐μm‐sectioned samples were placed on glass slides and stained with hematoxylin and eosin. Whole-slide scanning (20×) was performed on an EasyScan Infinity (Motic). ONL thickness was analyzed in ImageJ (NIH). Quantification of ONL thickness was performed in the medial retina. The areas analyzed were defined by distance from the optic disk. Ten measurements of ONL thickness were done per mouse.

Electron microscopy

Adult mice were perfused with 1× PBS followed by 4% paraformaldehyde (Electron Microscopy Sciences 15710) in PBS. Eyeballs were then carefully dissected and further fixed in 2.5% glutaraldehyde (Electron Microscopy Sciences 16200) and 2% paraformaldehyde in 0.1 M sodium cacodylate buffer (pH 7.4) for 1 hr at room temperature. Next, eyeballs were post-fixed in 1% OsO₄ for 1 hr at room temperature and en bloc stained with 2% aqueous uranyl acetate for 30 min. They were then dehydrated in a graded series of ethanol, going from 70% to 100%, and finally transferred to 100% propylene oxide before being embedded in EMbed 812 resin, polymerized at 60°C overnight. Samples from medial retinas were cut respectively into thin sections of 60 nm by a Leica ultramicrotome (UC7), placed on standard EM grids, and stained with 2% uranyl acetate and lead citrate. Retinal samples were examined with a FEI Tecnai transmission electron microscope at 80 kV accelerating voltage, and digital images were recorded with an Olympus Morada CCD camera and iTEM imaging software at the Yale Center for Cellular and Molecular Imaging (CCMI) Electron Microscopy Facility.

ERG recordings

All experimental animals were adapted in a dark room for 12 hr prior to recordings. Animals were anesthetized under dim red illumination using a 100 mg/kg ketamine and 10 mg/kg xylazine cocktail injected intraperitoneally and pupils were dilated by application of a 0.5% tropicamide eye drop (Sandoz 61214-354-01) at least 15 min before recordings. The cornea was intermittently irrigated with balanced salt solution to maintain the baseline recording and prevent keratopathy. Scotopic electroretinograms were acquired with UTAS ERG System with a BigShot Ganzfeld Stimulator (LKC Technologies, Inc). A needle reference electrode was placed under the skin of the back of the head, a ground electrode was attached subcutaneously to the tail, and a lens electrode was placed in contact with the central cornea. The scotopic response was recorded for different luminances (i.e., log₂ −2.1, –0.6, 0.4, and 1.4 cd.s/m²) using EMWin software, following the manufacturer’s instructions (LKC Technologies, Inc). The a-wave was measured as the difference in amplitude between baseline recording and the trough of the negative deflection, and the b-wave amplitude was measured from the trough of the a-wave to the peak of the ERG.

Tumor implantations

A total of 100,000 YUMM1.7 melanoma cells, resuspended in 50 ul of sterile 1× PBS, were subcutaneously injected into shaved rear flank of 6- to 12-week-old male mice. Mice were monitored for tumor growth by measuring the length and width of tumor masses using a caliper. Tumor volumes were scored with the formula (A × B²) × 0.4, in which A is the largest and B is the shortest dimension. Each mouse was said to have reached the end of its tumor-free survival when the largest dimension of its tumor was measured to be 5 mm. Mice were sacrificed once tumor growth reached an endpoint cutoff of 1000 mm³.

Anesthetized 8-to-12-week-old male mice were placed in a stereotactic apparatus and an incision was made with a scalpel over the cranial midline. A burr hole was made 1 mm lateral and 2 mm anterior to the bregma. A needle containing a suspension of GL261-luciferase (Luc) cells was inserted to a depth of 3 mm. After the needle is allowed to rest in the burr hole for 5 min, 10,000 GL261-Luc cells were infused over the course of 4 min. Once cells were injected, the needle was allowed to rest in the skull for 5 min before removal. Finally, the incision was closed with vetbond tissue adhesive and animals were administered the full course of postoperative analgesic drugs, according to regulatory guidelines and standards set by the Institutional Animal Care and Use Committee of Yale University. Intracranial tumor growth was monitored using the 3D image reconstruction feature of the IVIS Spectrum instrument. Mice received an intraperitoneal injection of 150 mg/kg of luciferin (Goldbio LUCK) prior to imaging. Tumor-bearing mice were checked daily for clinical signs of sickness behavior and were euthanized when one or more of the following symptoms were present: hunching, decreased activity, head tilt, weight loss, seizures, and failure to groom.

Generation of BMDCs

BMDCs were differentiated from granulocyte macrophage colony-stimulating factor (GM-CSF) as previously described (Inaba et al., 1992). Briefly, bone marrow progenitors were collected from the femurs of adult C57BL/6 WT mice and 10 × 10⁶ progenitor cells were cultured in RPMI media (Gibco 11875101) supplemented with 10% FBS (Sigma-Aldrich 18D078), 1% Penicillin-Streptomycin (Gibco 10378016), and 20 ng/ml of recombinant murine GM-CSF (PeproTech 315-03). On days 3 and 5, more supplemented media were added and cells were left in culture until day 7 when they would be ready to be used for generation of DC-Vax, as described below.

Preparation of dendritic cell vaccines

GL261 tumor cell lysates were made by subjecting cells to six rounds of rapid freeze–thaw cycles, comprised of 3 min of incubation in liquid nitrogen and 4 min of incubation at 56°C. BMDCs were then incubated with GL261 tumor lysate (1 mg of lysate/10 × 10⁶ BMDCs) for 2 hr at 37°C. After 2 hr, 1 ug/ml LPS (Sigma L2630) was added to BMDC-tumor lysate suspension and cells were incubated for 24 hr at 37°C. Next, supernatant was aspirated, and BMDCs were collected and washed with 1× PBS three times. Finally, 1 × 10⁶ GL261-lysate pulsed BMDCs were intraperitoneally injected into mice at days 14 and 21 post-intracranial implantation of GL261-Luc cells, as detailed above.

Total RNA isolation and sequencing analysis

RNA was collected from postnatal day 25 mice using a previously validated method (Xin-Zhao Wang et al., 2012). Briefly, after euthanasia, mice eyes were enucleated and the posterior eyecup was incubated on ice in 400 μl of RNAprotect (QIAGEN 76526) for 1 hr. RPE-containing tubes were agitated for 10 min to dislodge any RPE cells attached to the posterior eyecup and centrifuged for 5 min at 685 × g. The RPE pellet was then subjected to total RNA extraction using RNeasy Mini kit (QIAGEN 74106) following the manufacturer’s instructions. Similarly, total RNA from BMDMs was extracted from these cells using RNeasy Mini kit (QIAGEN 74106) following the manufacturer’s instructions.

RNA libraries from BMDMs and RPE cells were prepared at the Yale Keck Biotechnology Resource Laboratory from three to six biological replicates per condition. Samples were sequenced using 150 bp base pair paired-end reading on a NovaSeq 6000 instrument (Illumina). The raw reads were then subjected to trimming by btrim (Kong, 2011) to remove sequencing adaptors and low-quality regions. Next, reads were mapped to the mouse genome (GRCm38) using STAR (Dobin et al., 2013). Finally, the Deseq2 (Love et al., 2014) package was run to identify DEGs according to the p-values adjusted for multiple comparisons. Genes with p-adjusted values <0.05 and log₂ fold change ≤–1.25 or ≥1.25 were considered differentially expressed. All data RNA-sequencing datasets and the processed data that support the findings of this study have been deposited to the Gene Expression Omnibus (GEO) under accession ID: GSE205070.

Quantitative PCR analysis

Reverse transcription of RNA was performed utilizing iScript cDNA Synthesis Kit (Bio-Rad 1708891). Using KAPA SYBR Fast qPCR Kit (Kapa Biosystems KK4602), we amplified cDNA fragments and proceeded with qPCR reactions on CFX96 Thermal Cycler Real Time System (Bio-Rad). The reactions were normalized to three housekeeping genes (Gapdh, Hprt, and Rn18s), and specificity of the amplified products was verified by looking at the dissociation curves. All oligonucleotides for qPCR were either purchased from Sigma-Aldrich or produced at the Yale University Keck Oligonucleotide Synthesis Facility (see sequences in Appendix 1—table 1).

Statistical analysis

All statistical analyses were done using GraphPad Prism (GraphPad Software Inc). All data are shown as mean ± SEM, and each data point represents a unique animal. Statistical differences between experimental groups were determined by employing various tests, namely, Kaplan–Meier test, Mann–Whitney test, Student’s t-test, one-way and two-way ANOVAs. Additionally, hypergeometric distribution analysis was employed to determine which chromosomes are over-represented in the pool of DEGs associated with each genotype. The distribution of DEGs was said to be enriched on a chromosome when the probability of association with a chromosome had p-value <0.05 compared to the number of DEGs that would be expected to map to each chromosome by chance.

The authors acknowledge the members of the Rothlin-Ghosh laboratory for scientific discussions relating to the preparation of this manuscript. The authors also thank Dr Sagie Wagage for technical contributions with GL261 model experiments and Ms Veronica Galimberti for genotyping. This study was funded by NIH R01CA212376 (CVR and SG), Howard Hughes Medical Institute Faculty Scholar award (CVR) and a YSPORE Career Development Award DRP27 (CVR). SCF is supported by the Kim B and Stephen E Bepler Professorship in Biology. MA is the recipient of a long-term fellowship (BUIT 2012-5347) from the Dutch Cancer Society. LDH was awarded NSF DGE-1122492 and Richard K Gershon Fellowship. Research reported in this publication was supported by the National Cancer Institute of the National Institutes of Health under award number 2T32CA193200-06 (PIs PG and DFS).