Systematic Review of Biological Therapies for Atrial Fibrillation

Authors, Journal, Affiliations, Type, DOI

• Authors: Callum McRae, Anshul Kapoor, Pushpinder Kanda, Benjamin Hibbert, Darryl R. Davis
• Journal: Heart Rhythm, September 2019; 16(9):1399–1407
• Affiliations: Division of Cardiology, University of Ottawa Heart Institute, Ottawa, Ontario, Canada
• Type: Systematic review and meta-analysis of preclinical animal studies; CAMARADES and SYRCLE bias tools
• Funding: Canadian Institutes of Health Research (Clinician Scientist Award MC2-121291; Project Grant PJT-162088; graduate scholarships) — no industry funding
• DOI: https://doi.org/10.1016/j.hrthm.2019.03.021

Overview

This systematic review of 25 preclinical animal studies (screened from 3,877 PubMed + Embase articles; earliest published 2010) synthesises evidence for biological therapies — gene transfer, miRNA manipulation, and cell therapy — targeting the atrial substrate in AF. All 25 interventions reduced at least one primary outcome. Pooled meta-analysis showed biological therapies reduced AF inducibility by 85% (OR 0.15; 95% CI 0.07–0.35; P<0.01), reduced atrial scar burden by 6.7% (95% CI 4.2–9.2; P<0.01), and increased days in sinus rhythm by +6.4 (95% CI 5.83–6.97; P<0.01). Despite consistent preclinical efficacy, no biological therapy has entered clinical trials, limited by: delivery before AF onset (clinically unrealistic), invasive local delivery (non-scalable), transient transgene expression (adenovirus/plasmid), tiny animal group sizes (n=5–16), single-mechanism targeting, and publication bias. The review identifies AAV-based durable delivery and multi-target strategies as the translational path forward.

Keywords

Atrial fibrillation, animal models, cell therapy, gene therapy, biological therapy, atrial fibrosis, miRNA, connexin, SERCA2a, TGF-β

Key Takeaways

Study Selection and Scope

• Search: PubMed + Embase, July 16, 2018; terms: atrial fibrillation AND animal model; 3,877 articles → 25 included
• Inclusion criteria: Placebo-controlled animal model; biological therapy (not device or drug); recovered animals; quantification of AF incidence, atrial fibrosis, or electrophysiological parameters
• All 25 articles published since 2010; median journal impact factor 7.6±6.0; 24% in journals with IF ≥10
• Bias assessment: SYRCLE risk of bias tool — 10 low, 13 unknown, 2 high risk; large animal studies (8/13 low risk) outperformed small animal (2/12 low risk)
• Quality: CAMARADES checklist — no significant quality difference between large and small animal studies

Strategies and Studies (Table 1)

Disrupt cholinergic signaling (n=2):

• Aistrup: Plasmid-mediated Gαi/o inhibition → disrupted IK-ACh activation in dogs (carbachol + vagal induction)
• Zhang: miR-206 overexpression (lentivirus) → modulates intrinsic cardiac autonomic nerve remodeling via SOD1 in dogs

Increase atrial refractoriness (n=8):

• Amit, Liu, Soucek: Dominant-negative KCNH2 mutant (G627S/G628S; ↓IKr) — adenovirus, pigs (burst pacing)
• Lu: Anti-miR-328 (normalise ICaL) — plasmid, mice
• Luo: miR-26a overexpression (↓IK1) — adenovirus IV injection, mice; QT prolongation adverse event documented
• Jia: Anti-miR-1 (↓delayed rectifier) — lentivirus, rabbits
• Li: Anti-miR-328 (↑ICaL) — lentivirus, rabbits
• Lugenbiel: KCNK2 overexpression (↑TREK-1 current) — adenovirus, pigs
• Wang: SERCA2a overexpression (AAV9) — rabbit burst pacing model

Reduce apoptosis (n=1):

• Trappe: siRNA knockdown of caspase-3 — plasmid, pigs; +4.3 days in sinus rhythm

Reduce atrial fibrosis (n=10):

• Cardin, Adam, He, Huang: Anti-miR-21 — multiple species (rat/mouse/rabbit); 4 independent labs — most replicated strategy; fibrosis reduced 9.5% (Cardin), 4.9% (Huang)
• Fan, Zhou ×2: ACE2 overexpression — adenovirus epicardial, dogs; 3 studies by Zhou group; fibrosis −8.5% (Fan), −6.4% (Zhou 2015)
• Kunamalla: Dominant-negative TGF-β1 receptor — plasmid, dog HF model; fibrosis −8.7%
• Xu: miR-30c overexpression (AAV) — rat aortic constriction; fibrosis −10.4%
• Wang: miR-27b overexpression (AAV, pericardial) — mouse angiotensin II; fibrosis −15.7%

Increase conduction velocity (n=2):

• Igarashi: GJA1 (Cx43) or GJA5 (Cx40) overexpression — adenovirus epicardial, pigs; preserved conduction velocity
• Bikou: GJA1 (Cx43) overexpression — adenovirus, pigs; +6.6 days in sinus rhythm; validated by independent lab

Cell therapy (n=1):

• Ramos: Mesenchymal stem cell IV infusion — rat sleep apnea model; fibrosis −2.4%; only cell therapy study

Pooled Meta-Analysis Results (Table 3, Figure 3)

• Days in sinus rhythm (3 studies; continuous): Treatment +6.4 days vs control (95% CI 5.83–6.97; P<0.01)
• AF inducibility (8 studies; dichotomous): OR 0.15 (95% CI 0.07–0.35; P<0.01) — 85% reduction in likelihood of inducing AF
• Atrial fibrosis (8 studies; continuous): −6.7% absolute fibrosis content (95% CI 4.2–9.2; P<0.01)
• No study reported a failed intervention (primary outcome unaffected) → likely reflects strong publication bias
• Only 1 study documented an adverse event: QT prolongation after IV adenovirus (miR-26a)
• 4 strategies validated by ≥2 independent laboratories: anti-miR-21 (4 labs), KCNH2 mutant (2 labs), miR-328 suppression (2 labs), GJA1 overexpression (2 labs)
• 1 dose-response relationship established (Aistrup only)

Why No Biological Therapy Has Reached Clinical Trials

1. Pre-AF delivery only: All 25 studies delivered therapy before AF development — there is no clinically realistic scenario to identify and treat patients before AF onset; the "therapy for established AF" paradigm is absent from the preclinical literature
2. Invasive local delivery: 80% of studies used intramyocardial or epicardial direct injection — clinically non-scalable; systemic IV delivery available for some constructs but with off-target risk
3. Transient expression: Adenovirus/plasmid provide robust expression for only 3–4 days; only 2 studies followed animals beyond 21 days; lasting expression requires AAV (few studies) or permanent gene editing
4. Single-mechanism targeting: All studies targeted one pathway; AF is multifactorial — heart failure AF has autonomic hyperactivity + fibrosis + ion channel remodeling simultaneously; "jack of all trades" single agent likely insufficient
5. Tiny group sizes: n=5–16 animals per treatment group; only 3 studies had >10/group — severely limits statistical power and replicability
6. Publication bias: Zero failed studies published; no dose-response in 24/25 studies; SYRCLE flagged 2 studies as high risk (animal number inconsistencies)
7. Cost and GMP manufacturing: Mouse doses are 1,000-fold smaller than clinical doses; Good Manufacturing Practices-quality xenogen-free manufacturing is cost-prohibitive for each dose

Emerging/Next-Generation Targets Not Yet Explored

• Atrial inflammation (lymphomononuclear infiltration in lone AF biopsies)
• Autonomic tone (β-adrenoceptor desensitization in HF+AF)
• Circadian variations in ion channel transcription
• Mechanoelectric feedback
• Hormonal modulation (AT1R-Ca²⁺, adrenergic)
• Multi-target approaches at regulatory network nodal points

Comparison to Successful Biological Therapies in Other Domains

• Successful clinical biological therapies (transplant, immunization, antigen-targeted immune cells — CAR-T) all produce lasting effects without long-term repetitive treatment
• AF biological therapies in this review are "transient modification of a single transcript" — more similar to a drug than a true biological; cannot be repeatedly re-administered (virus/plasmid) without systemic off-target effects
• Tissue-specific AAV vectors with asialoglycoprotein-like receptor targeting (analogous to ASO/PCSK9 targeting in liver) offer a potential path to systemic atrial-selective delivery

Limitations of the Document

• All evidence preclinical: Zero human data; animal-to-human translation has historically failed for most cardiac therapies
• Publication bias confirmed: No failed study in 25 published; true negative studies likely exist in file-drawer
• Short follow-up: Only 2 studies >21 days; biological durability is untested beyond 3 weeks in 23/25 studies
• Tiny sample sizes: n=5–16 per group; most studies severely underpowered for detecting modest effects or safety signals
• Heterogeneous endpoints: AF inducibility, spontaneous AF burden, atrial fibrosis, and days in sinus rhythm are not uniformly comparable across animal models and induction methods
• No dose-response in 24/25 studies: Cannot determine efficacy thresholds or safety margins
• Delivery systems not clinically compatible: 80% direct intramyocardial/epicardial; no validated systemic targeting approach for atrial tissue
• No spontaneous AF models in most studies: Burst pacing-induced AF is the dominant model; spontaneous AF models (recently developed) better mimic human disease but were not used in most included studies
• Industry-independent CIHR funding is a strength for bias; but single-group focus (Zhou group: 3 ACE2 studies = 12% of literature) represents concentration risk
• Systematic review only up to 2018: Significant preclinical progress (AAV-based strategies, CRISPR-based editing) has likely occurred since this review

Key Concepts Mentioned

• concepts/AF-Biological-Therapy — the full framework of preclinical biological strategies
• concepts/AAV-Gene-Delivery — AAV9/AAV-based durable delivery as translational path forward
• concepts/Gene-Silencing-Therapy — miRNA suppression/overexpression strategies (anti-miR-21, miR-26a, miR-328)
• concepts/Electrical-Remodeling — ionic targets of AF biological therapies (ICaL, IKr, IK1, IKACh)

Key Entities Mentioned

• entities/Atrial-Fibrillation — disease target of all 25 preclinical studies

Wiki Pages Updated

• wiki/sources/biological-tx-af-hrs-2019.md — created (this file)
• wiki/concepts/AF-Biological-Therapy.md — created
• wiki/entities/Atrial-Fibrillation.md — Gene Therapy section updated with meta-analytic data; source_count 22→23; source and concept link added
• wiki/sourceindex.md — new entry added
• wiki/wikiindex.md — new concept entry; AF entity updated
• log.md — ingest entry prepended