precision-medicine-mrna-design

# Precision Medicine & mRNA Design ## Overview Precision veterinary medicine tailors cancer therapy to individual tumor genetics and patient pharmacogenomics. This skill covers three pillars: **(1) mRNA neoantigen vaccine design** for personalized immunotherapy, **(2) pharmacogenomics** for optimal drug dosing and efficacy prediction, and **(3) companion diagnostics** linking genetic markers to therapeutic response. > **⚠️ Note on the worked examples in this skill.** All example values below — sequencing metrics, specific mutations, epitope predictions, mRNA/LNP formulations, immune-monitoring numbers, and clinical outcomes — are **synthetic and illustrative**. They are teaching values chosen to show the *shape* of each step. They are **not real patient data** and are **not** derived from any specific animal, including the real "Rosie" case. For what is actually publicly known (and, importantly, what is *not* confirmed) about the real Rosie case, see [`cases/rosie-2026`](../../../cases/rosie-2026/README.md). VetClaw was not involved in that case. ## When to Use - User designs personalized neoantigen mRNA vaccine for canine/feline cancer patient - User interprets tumor genomic sequencing (somatic mutations, tumor mutational burden) for treatment selection - User predicts drug response/toxicity based on pharmacogenomic variants (MDR1, CYP450) - User selects targeted therapy matching tumor driver mutations (BRAF, KIT, TP53) - User understands mRNA construct design (5' cap, ORF, poly-A tail, delivery logistics) - Keywords: precision medicine, mRNA vaccine, neoantigen, pharmacogenomics, targeted therapy, companion diagnostics, personalized immunotherapy - **Related skill:** For the detailed 8-step neoantigen vaccine pipeline (sequencing, variant calling, DLA typing, prediction, synthesis), see `neoantigen-vaccine-design`. This skill covers the broader precision medicine framework; that skill covers the specific mRNA vaccine bioinformatics workflow. ## The Precision Medicine Framework **Three Pillars of Personalization:** ``` ┌─────────────────────────────────────┐ │ PRECISION VETERINARY MEDICINE │ ├─────────────────────────────────────┤ │ │ │ 1. MRNA NEOANTIGEN VACCINES │ │ └─ Tumor sequencing │ │ └─ Neoantigen prediction │ │ └─ mRNA synthesis │ │ └─ Immunotherapy │ │ │ │ 2. PHARMACOGENOMICS │ │ └─ Patient germline variants │ │ └─ Drug metabolism prediction │ │ └─ Dosing optimization │ │ └─ Toxicity prediction │ │ │ │ 3. COMPANION DIAGNOSTICS │ │ └─ Tumor driver mutations │ │ └─ Predictive biomarkers │ │ └─ Therapy selection │ │ └─ Prognosis assessment │ │ │ └─────────────────────────────────────┘ ``` ## Pillar 1: mRNA Neoantigen Vaccine Design ### What is a Neoantigen? **Definition:** Tumor-specific mutation that creates a novel protein epitope not present in normal tissue. Neoantigens are immunogenic (unlike wild-type self-antigens); vaccines targeting neoantigens activate tumor-destroying CD8+ T cells without autoimmunity risk. **Example: TP53 Mutation in Canine Osteosarcoma** ``` Wild-type TP53 sequence (normal cell): MPPQPQ...SVQLG... (normal p53 protein) Tumor TP53 mutation (R175H): MPPQPQ...SHHQG... (mutant p53 protein - neoantigen!) Neoantigen epitope (8-10 amino acid segment): ...HHQG... ← differs from wild-type; recognized as foreign by immune system mRNA vaccine presents this epitope → CD8+ T cells attack tumor cells expressing R175H ``` ### Workflow: Tumor Sequencing to mRNA Vaccine **Step 1: Tumor Tissue Collection & Sequencing** 1. **Surgical biopsy:** Fresh tumor sample (optimal) or archived tissue 2. **DNA extraction:** Isolate tumor cell DNA 3. **Matched normal tissue:** Peripheral blood or normal tissue (distinguish somatic vs. germline mutations) 4. **Sequencing strategy:** - **Whole Exome Sequencing (WES):** Cost-effective (~$1,000-3,000); covers coding regions + splice sites - **Whole Genome Sequencing (WGS):** Complete coverage (~$5,000-10,000); captures intergenic, deep intronic mutations - **Targeted panels:** Deep coverage of known driver genes (BRAF, KIT, TP53, etc.); cost-effective if mutations well-characterized 5. **Depth requirements:** ≥100x coverage (tumor), ≥30x coverage (normal) for confident variant calling **Illustrative Example: Canine Mast Cell Tumor Sequencing (synthetic data — not a real patient)** ``` Patient: illustrative canine patient with aggressive mast cell tumor Sample: Fresh tumor tissue (surgical excision) Normal comparison: Peripheral blood leukocytes Sequencing: WES (Illumina NovaSeq; 150x coverage) Tumor mutational burden: 8.3 mutations/Mb (moderately high) Key somatic mutations identified: TP53 R248Q, PTEN loss, NRAS Q61R Germline variants: None pathogenic (clear breeding risk) ``` **Step 2: Mutation Calling & Annotation** 1. **Align reads:** Map sequencing reads to reference genome (CanFam3.1 for dogs) 2. **Call variants:** Identify somatic mutations (tumor) vs. germline (normal) 3. **Annotate mutations:** - Functional impact: Frameshift, missense (conservative vs. deleterious), nonsense - Cancer-related databases: ClinVar, COSMIC, OncoKB - Conservation: Is position evolutionarily conserved? (suggests functional importance) 4. **Filter variants:** Remove sequencing artifacts, common SNPs (>1% population frequency) 5. **Prioritize mutations:** Focus on high-impact mutations (driver genes) **Step 3: Neoantigen Prediction** 1. **In silico prediction:** Use algorithms to identify peptide epitopes likely recognized by MHC (Major Histocompatibility Complex) - **Algorithm examples:** - NetMHC (MHC binding affinity prediction) - MixMHCpred (peptide presentation) - DeepImmuno (deep learning epitope prediction) - **Input:** Mutant amino acid sequence, canine MHC type (dog MHC homologous to human HLA) 2. **Scoring criteria:** - Strong binder: MHC binding affinity <500 nM (high probability of presentation) - Wild-type comparison: Mutant should bind better than wild-type (selectivity) - Multiple epitopes: Predict 8-mer, 9-mer, 10-mer peptides (different HLA types) 3. **Machine learning refinement (Optional):** - Use LLMs or AlphaFold to predict 3D structure of epitope-MHC complex - Higher predicted stability = more likely immunogenic - Example: AlphaFold2 structure prediction of mutant peptide in MHC binding groove **Illustrative Example: Neoantigen Prediction for a TP53 R248Q Mutation (synthetic data)** ``` Mutation: TP53 R248Q (Arginine → Glutamine at position 248) Wild-type epitope: ...LSPPQK|RQSLP... Mutant epitope: ...LSPPQK|QQSLP... (R248Q) Predicted MHC-binding epitopes: Epitope 1: QRQSLPGV (8-mer) - Binding affinity 0.3 µM (strong binder) Epitope 2: QRQSLPGVG (9-mer) - Binding affinity 0.5 µM (strong binder) Epitope 3: KQQSLPGVG (9-mer) - Binding affinity 2.1 µM (moderate binder) Selected for vaccine: Top 2-5 epitopes with strongest binding affinity + wild-type selectivity Final count: 5 neoantigen epitopes selected for mRNA construct ``` **Step 4: mRNA Construct Design** mRNA vaccine structure (5' → 3'): ``` 5' CAP ─ UTR5 ─ ORF (Neoantigen Codons) ─ UTR3 ─ PolyA tail ─ 3' ↑ ↑ ↑ ↑ Ribosome Protein Stability mRNA binding synthesis signal tail ``` **Components:** 1. **5' Cap (m7G):** Protects from exonuclease digestion; recognized by translation machinery - Structure: 7-methylguanosine linked via unusual 5'-5' triphosphate bond - Function: Facilitates ribosome binding; reduces innate immune activation 2. **5' UTR (Untranslated Region):** 50-150 nucleotides - Contains ribosome binding site (Kozak sequence) - Optimized for translation efficiency (high GC content improves stability) 3. **ORF (Open Reading Frame):** Encodes neoantigen protein - Typically 600-1,500 bp (200-500 amino acids) - Multiple neoantigens concatenated (illustrative construct: 5 neoantigens fused with linkers) - **Codon optimization:** Synonymous substitutions for faster translation - Example: Replace rare codons (CGA for Arg) with common codons (CGC) - Improves ribosome speed, protein production - **Signal peptide prepended:** Optional; directs protein to endoplasmic reticulum for cross-presentation to CD8+ T cells 4. **Spacer/Linker Sequences:** Between neoantigen epitopes - Function: Prevent fusion artifacts, allow independent cleavage - Design: Protease recognition sites (e.g., furin cleavage sites) for natural processing 5. **3' UTR (Untranslated Region):** 50-200 nucleotides - Contains mRNA stability elements (AAUAAA polyadenylation signal recognized by nuclear machinery) - Optimized for half-life (typically 2-4 hours in transfected cells) 6. **Poly-A Tail:** 100-250 adenine nucleotides - Protects 3' end from degradation - Enhances translation efficiency - Recognized by poly-A binding proteins (PABP) **Example mRNA Construct (Synthetic Design):** ``` 5' m7G-GCCGCCACCAUGGCCAUGGCGCGCUUUGAGCCAUGCGC ↑ ↑ 5' cap Kozak (start codon) CGCGAAAAGACUAUAAACUGCUAGCGAAAA[NEOANTIGEN1]GGGCUGCGAA AAG[NEOANTIGEN2]CGCUUACGAGCUAA[NEOANTIGEN3]... [5 neoantigens concatenated with furin cleavage sites] ...AAUAAAGGGGAAAA[A]100 3' ↑ ↑ poly-A signal poly-A tail ``` **Step 5: mRNA Synthesis & Quality Control** 1. **In vitro transcription (IVT):** - Template DNA with promoter (T7 RNA polymerase) → mRNA transcript - One-pot reaction: DNA template + NTPs + polymerase → linear mRNA - **Capping enzyme:** Add m7G cap co-transcriptionally (cap-0) or post-transcriptionally (cap-1) - **Polyadenylation:** Add poly-A tail enzymatically (poly-A polymerase) or encoded in template 2. **Purification:** - RNeasy column: Remove proteins, salts, free nucleotides - HPLC (optional): Polishing for clinical-grade purity - Precipitate with LiCl or isopropanol; resuspend in RNase-free buffer 3. **Quality control:** - **Integrity:** Agarose gel (single band = full-length mRNA; degradation = smearing) - **Concentration:** Nanodrop (A260/280 ratio ~1.8-2.0 = pure RNA) - **Endotoxin:** LAL assay (<0.5 EU/µg required for clinical use) - **Sterility:** Bacterial/fungal culture (48-72 hr growth assay) - **Functionality:** Transfect mammalian cells, measure protein expression by Western blot/ELISA **Illustrative Quality Metrics (synthetic data):** ``` Construct: 2,847 bp (5 concatenated neoantigens) Yield: 850 µg from 10 mL IVT reaction (4.25 mg/mL) Integrity: 96% full-length (gel analysis) Endotoxin: 0.08 EU/µg (well below clinical threshold) Sterility: Negative (no growth at 48 hrs) Protein expression: 450 ng/mL neoantigen protein (HEK293T cells, 24 hrs post-transfection) ``` ### Step 6: mRNA Formulation (Lipid Nanoparticles) **Challenge:** Naked mRNA is degraded rapidly (half-life <5 minutes in serum); immune-stimulating (dsRNA triggers TLR3) **Solution:** Encapsulate in Lipid Nanoparticles (LNPs) **LNP Composition (4-component system):** | Component | Function | Example | |-----------|----------|---------| | Ionizable Lipid | mRNA binding, cellular uptake | SM-102, mRNA-1273 lipid ionizable component | | Structural Lipid | Particle scaffold | DSPC (1,2-distearoyl-sn-glycero-3-phosphocholine) | | PEG-Lipid | Surface coating, circulation time | PEG2000-DMG (reduces opsonization) | | Cholesterol | Membrane fluidity | 20-40% of lipid composition | **LNP Biophysics:** - Particle size: 80-120 nm (optimal for lymphoid tissue drainage) - Surface charge: Slightly positive (enhances cellular uptake) - PEG density: ~2% weight (stealth effect; prolongs half-life from minutes to hours) - Endocytosis pathway: Clathrin-mediated; trafficking to early endosome → release into cytoplasm (pH-dependent ionizable lipid) **LNP Preparation (Self-Assembly):** ``` Step 1: Mix lipid stock solutions in ethanol: - Ionizable lipid: 50% molar ratio - Structural lipid: 38.5% - Cholesterol: 10% - PEG-lipid: 1.5% Total volume: 500 µL ethanol Step 2: Dilute mRNA in 50 mM sodium acetate pH 4.0 (aqueous phase) Step 3: Rapid mixing (microfluidic mixer or syringe injection): - Ethanol lipid stream meets aqueous mRNA stream - Lipids self-assemble around mRNA (rapid hydration) - Nanoparticles form spontaneously (milliseconds) Step 4: Buffer exchange: - Dialyze against PBS or saline (remove ethanol, neutralize pH) - Remove free mRNA (0.5-2% typically remains unencapsulated) Step 5: Concentration: - Vivaspin or tangential flow filtration - Target: 1-5 mg mRNA/mL LNP suspension - Sterile filtration (0.22 µm) for clinical use ``` **Quality Metrics (Post-Formulation):** - **Encapsulation efficiency:** >85% (HPLC or qRT-PCR of free mRNA) - **Particle size:** 95 ± 10 nm (dynamic light scattering) - **Polydispersity index (PDI):** <0.2 (uniform size distribution) - **mRNA:Lipid ratio:** Typically 1:20 to 1:40 (w/w) - **Endotoxin:** <0.5 EU/µg (limulus amebocyte lysate test) - **Sterility:** Negative 48-hr culture **Illustrative Example (synthetic LNP formulation):** ``` mRNA: 850 µg (pure, full-length neoantigen vaccine) Ionizable lipid (SM-102): 42.5 mg Structural lipid (DSPC): 32.2 mg Cholesterol: 8.5 mg PEG-lipid (PEG2K-DMG): 1.3 mg Final LNP concentration: 2.8 mg mRNA/mL Encapsulation: 92.1% (8.2 µg free mRNA recovered) Particle size: 98 ± 7 nm Endotoxin: 0.12 EU/µg (acceptable) Dose per injection: 100 µg mRNA (35.7 µL LNP suspension) Formulation: Suspension in sterile saline; stored at -20°C Stability: >6 months frozen; <4 hrs at room temperature ``` ### Step 7: Vaccine Administration & Immune Monitoring **Administration (Veterinary Setting):** - **Route:** Subcutaneous or intramuscular injection - **Schedule:** Typically 3-5 doses, 2-week intervals (priming + booster) - **Dose:** 50-100 µg mRNA per injection (species-adjusted) - **Site:** Lateral thorax or hindlimb (allows drainage to regional lymph nodes) **Immune Responses Measured:** 1. **CD8+ T-cell Response (Primary):** - IFN-γ ELISPOT: Measure T cells secreting interferon-gamma upon neoantigen restimulation - Flow cytometry: CD3+, CD8+, tetramer+ cells (tetramer = neoantigen-MHC complex) - Target: >100 IFN-γ-secreting cells per 10^6 PBMCs post-vaccination 2. **CD4+ Helper T-cell Response:** - IL-2, IL-4, IL-17 secretion (polyfunctional response desired) - Th1 bias preferred for anti-tumor efficacy 3. **Antibody Response:** - Anti-neoantigen IgG ELISA (typically weak with mRNA; not primary response) 4. **Neoantigen-Specific T-cell Clones:** - TCR sequencing: Identify expanded T-cell clones recognizing vaccine epitopes - Functional testing: Isolate expanded clones, measure cytotoxicity against tumor cells **Illustrative Example (synthetic immune-response data):** ``` Timeline: Pre-vaccine → 1 week post-dose 1 → 1 week post-dose 3 Pre-vaccine (Baseline): - IFN-γ ELISPOT (neoantigen restimulation): 2 cells/10^6 PBMCs (background) - CD8+ tetramer+ cells: <0.1% (undetectable) Post-dose 3 (Week 6): - IFN-γ ELISPOT: 285 cells/10^6 PBMCs (142-fold expansion!) - CD8+ tetramer+ cells: 2.3% of CD8+ T cells - CD8+ subset: Predominantly memory phenotype (CD45RA-, CCR7-); TEM = tissue-resident - TCR clonality: 15 dominant clones identified; 3 clones account for 45% of response - Polyfunctionality: 60% of tetramer+ cells produce IFN-γ, 35% produce TNF-α Functional Activity: - Isolated neoantigen-specific CD8+ T cells co-cultured with autologous tumor cells - Cytotoxicity: 35% specific lysis (4-hour 51Cr release assay) - IFN-γ secretion: 450 pg/mL (positive control) Assessment: Strong, polyfunctional CD8+ T-cell response; suitable for therapy ``` ### Step 8: Clinical Outcomes & Translational Data **Illustrative Clinical Course (synthetic example — not a real patient):** ``` Pre-treatment: - Diagnosis: Mast cell cancer, confirmed by histopathology - Staging: Abdominal ultrasound, thoracic radiographs (no distant metastases) - Prognosis: ~5-month median survival without treatment - Performance: ECOG 1 (mild activity reduction) Treatment: - Surgical excision (Day 0): Primary tumor removed with margins - Pathology: High-grade mast cell tumor, high mitotic index - mRNA vaccine (Days 10, 24, 38): 3-dose series, 100 µg each - Doxorubicin chemotherapy (Days 7, 21, 35, 49): Adjuvant 30 mg/m2 IV Monitoring: - Immune response: Robust CD8+ T-cell priming (as above) - Imaging: Abdominal ultrasound @ weeks 4, 8, 12 (no new lesions) - Bloodwork: CBC normal; chemistry panel normal; no organ toxicity - Quality of life: Normal appetite, exercise tolerance, no pain Outcome (illustrative — hypothetical values, not a real result): - Overall survival: illustrative endpoint only (no real outcome is claimed here) - Disease-free interval: illustrative only - Historical-control context (general): untreated high-grade MCT often carries a guarded prognosis; adjuvant chemotherapy may extend it — consult current oncology references - Translational relevance: comparative-oncology data of this kind can, in principle, inform human mRNA-vaccine trial design ``` ## Pillar 2: Pharmacogenomics ### Concept: Why Genetics Matter for Drug Response