Anne-Mette Bjerregaard

MuPeXI: prediction of neo-epitopes from tumor sequencing data

Personalization of immunotherapies such as cancer vaccines and adoptive T cell therapy depends on identification of patient-specific neo-epitopes that can be specifically targeted. MuPeXI, the mutant peptide extractor and informer, is a program to identify tumor-specific peptides and assess their potential to be neo-epitopes. The program input is a file with somatic mutation calls, a list of HLA types, and optionally a gene expression profile. The output is a table with all tumor-specific peptides derived from nucleotide substitutions, insertions, and deletions, along with comprehensive annotation, including HLA binding and similarity to normal peptides. The peptides are sorted according to a priority score which is intended to roughly predict immunogenicity. We applied MuPeXI to three tumors for which predicted MHC-binding peptides had been screened for T cell reactivity, and found that MuPeXI was able to prioritize immunogenic peptides with an area under the curve of 0.63. Compared to other available tools, MuPeXI provides more information and is easier to use. MuPeXI is available as stand-alone software and as a web server at http://www.cbs.dtu.dk/services/MuPeXI.

An Analysis of Natural T Cell Responses to Predicted Tumor Neoepitopes

Personalization of cancer immunotherapies such as therapeutic vaccines and adoptive T-cell therapy may benefit from efficient identification and targeting of patient-specific neoepitopes. However, current neoepitope prediction methods based on sequencing and predictions of epitope processing and presentation result in a low rate of validation, suggesting that the determinants of peptide immunogenicity are not well understood. We gathered published data on human neopeptides originating from single amino acid substitutions for which T cell reactivity had been experimentally tested, including both immunogenic and non-immunogenic neopeptides. Out of 1,948 neopeptide-HLA (human leukocyte antigen) combinations from 13 publications, 53 were reported to elicit a T cell response. From these data, we found an enrichment for responses among peptides of length 9. Even though the peptides had been pre-selected based on presumed likelihood of being immunogenic, we found using NetMHCpan-4.0 that immunogenic neopeptides were predicted to bind significantly more strongly to HLA compared to non-immunogenic peptides. Investigation of the HLA binding strength of the immunogenic peptides revealed that the vast majority (96%) shared very strong predicted binding to HLA and that the binding strength was comparable to that observed for pathogen-derived epitopes. Finally, we found that neopeptide dissimilarity to self is a predictor of immunogenicity in situations where neo- and normal peptides share comparable predicted binding strength. In conclusion, these results suggest new strategies for prioritization of mutated peptides, but new data will be needed to confirm their value.

Corrigendum: An Analysis of Natural T Cell Responses to Predicted Tumor Neoepitopes

[This corrects the article on p. 1566 in vol. 8, PMID: 29187854.].

Prediction of neoepitopes from murine sequencing data

We recently published the tool MuPeXI, the mutant peptide extractor and informer, enabling neoepitope prediction from tumor sequencing data [1]. MuPeXI is originally designed for variant calls obtained from sequencing data of human origin but increasing interest to determine neoepitopes in murine models have encouraged us to update and test MuPeXI for mouse compatibility. The murine-compatible MuPeXI is now available as a command line tool (https :// githu b.com/ambj/MuPeX I) together with a mouse-specific web server (http://www.cbs.dtu.dk/servi ces/MuPeX I-mouse /). Despite the interest for determining neoepitopes from preclinical mouse models, only few tools for neoepitope prediction have been designed and evaluated to allow neoepitope prediction from data of murine origin. To fulfill this need, we optimized MuPeXI to enable identification of murine neopeptides. MuPeXI is now compatible with the genetic reference of mus musculus, as well as the two commonly used mouse strains, BALBc and C57BL/6. To test the NGS pipeline and optimize MuPeXI, we evaluate the neoepitope landscape in the CT26 tumor cell line, which has been extensively used in mode-of-action studies in syngeneic mouse tumor models [2], and proven especially valuable as an experimental model for immune therapy interventions. We used sequencing data from Castle et al. [2] and Mosely et al. [3], including both the CT26 cell line (CL) and the CT26 tumors grown in vivo on BALBc mice (TU). The NGS analysis pipeline was followed as suggested by Genome Analysis Tool Kit (GATK), best practice guidelines, using the same tools as in the original MuPeXI paper [1]. References were downloaded from Ensembl’s mouse genome assemble, further detail can be found in the MuPeXI user manual (https :// githu b.com/ambj/MuPeX I/blob/maste r/doc/MuPeX I_User_ Manua l.md#refer ences ). In the analysis we incorporated the new binding predictor netH2pan into MuPeXI. NetH2pan is trained solely on mouse-binding affinity and eluted ligand data [4], thereby providing the most suitable H2-binding predictions for the neopeptides extracted from somatic variant of murine sequencing data by MuPeXI. The mutational landscape was compared for the three samples tested, and although the total number of mutations identified in the three samples were high, it was in accordance with the original papers. The analysis revealed that only a fraction of the mutations, 7898 (~ 26% of the total) was identified in all 3 samples (Fig. 1a). The number of nonsynonymous mutations (NSmut) did not vary substantially compared to the total number of mutations identified, i.e., Castle: 5994–3.8%, Mosely CL: 4921–3.1%, Mosely TU: 4793–3.0% (Fig. 1b). Besides missense variant (MV) mutations, frame shifts, and indels (FI) were also identified, but present to a lesser degree (MV: 15349–9.6%, FI: 359–0.2%). To identify how many of the NSmuts lead to potential neoepitopes, the updated murine-compatible MuPeXI software was used with the relevant BALBc references. We identified 79,958 (61% of all) neopeptides shared among the two tumor samples obtained from two different studies and the original CT26 cell line sample (Fig. 1c). Of these, 4399 (3.4%) had an eluted ligand percentile rank score (%Rank EL) below 2 and were considered binding peptides in all 3 samples (Fig. 1c, blue). A total of 7034 (5.4%) peptides were identified as binders and therefore potential neoepitopes out of the total 124,467 unique neopeptides extracted from all 3 samples. Of the binding peptides, 61 potential minimal neopeptides matched previously described long peptides shown to evoke immune responses [5]. The potential neoepitopes originate from various types of NSmuts, including Anne-Mette Bjerregaard and Thomas Kainamura Pedersen contributed equally.

Evaluating prediction strategies for identification of T cell responsive mutation-derived neoepitopes in cancer

CD4+Foxp3+ regulatory T cells (Tregs) are the main regulators of peripheral tolerance and prevent the development of fatal autoimmune disease in humans and mice. Furthermore, Tregs have also been implicated in suppressing anti-tumour immune responses and are often enriched at sites of primary and metastatic tumours. While studies have shown the effect of Treg ablation on the control of primary tumours, few studies have examined their contribution to metastasis progression. In this thesis I hypothesised that the depletion of Tregs could promote control over metastasis. To address this, a highly metastatic murine mammary carcinoma cell line 4T1 was injected into transgenic mice expressing the diphtheria toxin receptor in Foxp3+ cells. Foxp3+ cells were depleted by administration of diphtheria toxin and the impact of this on growth of primary tumours and metastases was assessed and measured in vitro clonogenic assays. Results of these experiments indicated that Tregdepletion led to control of primary tumour growth and in some mice to control of metastases. Control of metastases was linked to control of primary tumour growth. In order to measure metastasis in vivo, a PET/CT imaging technique was optimized. Primary tumours and large metastatic nodules were successfully imaged in mice using F18 FDG as a radiotracer. However, the studies described herein revealed that micrometastases in mouse lungs were too small to be reliably identified using PET data parameters. CT imaging did however enable detection of increases in tissue density within the lungs, which was suggestive of micrometastases. Data obtained in this way also indicated that Treg-depletion promotes control of metastasis in some mice. Collectively, the findings described in this thesis indicate that Tregdepletion can contribute to control of metastatic disease and should therefore represent an important component of novel immunotherapies.Changes in microbiome, mucosal immunity and intestinal integrity have been associated with the onset of Type 1 Diabetes (T1D) in children. Toll-like Receptors (TLR) have been associated all three factors. The role of TLR and their effects on microbiome in autoimmunity were studied by crossing TLR1,2,4,6,9 and MyD88 targeted deficiency mutations to the type 1 diabetes (T1D)-prone NOD mouse strain. While NOD.Tlr9-/- and NOD.Tlr6-/- mice were unaffected, T1D in NOD.Tlr4-/- and NOD.Tlr1-/- mice was exacerbated and that in NOD.Myd88-/- and NOD.Tlr2-/- mice ameliorated. Physical parameters of the intestines were compared; ileal weight was reduced in NOD.Tlr1-/-mice. Similarly, by histology, these mice had reduced villus length and width. The intestinal microbiomes of NOD wild-type (WT), NOD.Tlr1-/-, NOD.Tlr2-/- and NOD.Tlr4-/- mice were compared by high throughput sequencing of 16S ribosomal DNA (rDNA), in two cohorts, 18 months apart. Analysis of caecal 16S sequences clearly resolved the mouse lines and there were significant differences in beta diversity between the strains, with individual bacterial species contributing greatly to the differences in the microbiota of individual TLR-deficient strains. To test the relationship between microbiome and T1D, all strains were re-derived onto the parental NOD/Lt line and the incidence of T1D re-assessed within two generations. All rederived lines expressed an incidence of disease similar to that of the parental line. TLR deficiencies are associated with changes in microbiome; changes of microbiome are associated with T1D; the effects of TLR deficiencies on T1D appear to be mediated by their effects on gut flora.Intestinal TCRb+CD4-CD8b-CD8a+ (CD8aa) IELs alleviate T cell induced colitis and have been suggested to play a role in virus infection and cancer. Their thymic development has been elucidated to some extent, as IEL precursors (IELp) are known to be CD4-CD8-CD5+TCRb+, but is not yet fully understood. Within the thymus, mature IELp were identified based on their expression of CD122 and MHC class I. Two major phenotypic subsets exist within this mature thymic IELp population: a PD1+Tbet- population that preferentially expresses a4b7, and a PD1-Tbet+ population with preferential CD103 expression. These two populations were also distinct in their Valpha repertoire. The PD1+a4b7+ population contains clones that are strongly self-reactive as judged by Nur77GFP and their dramatic increase in Bim deficient mice, while the PD1-Tbet+ population did not show these characteristics. Both gave rise to CD8aa IELs upon adoptive transfer into RAG-/- recipients. However intrathymic labeling revealed that PD1+a4b7+ IELp were the major thymic emigrating population, and emigration was S1P1-dependent. In contrast, PD1-Tbet+ IELp expressed CXCR3, were retained, and accumulated in the thymus with age. Preliminary immunofluorescence data furthermore indicate differential thymic cortico-medullary localization of the IELp subtypes. These experiments more precisely define the behavior of IEL precursors.