Progressive supranuclear palsy (PSP) has traditionally been considered a relatively uniform four-repeat tauopathy. However, emerging evidence from Forrest et al. (2026) demonstrates that HLA (human leukocyte antigen) haplotypes correlate with distinct clinical and immunopathological phenotypes, suggesting that PSP may encompass multiple aetiological-pathogenic events including targetable autoimmune mechanisms [1]. This finding has profound implications for patient stratification in disease-modifying therapy trials.
HLA molecules are critical for antigen presentation to T cells and regulate immune responses. Prior studies identified rare HLA haplotypes in PSP, but the functional consequences of these associations remained unknown [2]. The 2026 study bridges this gap by combining neuropathological analysis with machine learning to demonstrate that HLA-defined groups show distinct neuroinflammatory profiles and symptom progression sequences.
HLA genes encode major histocompatibility complex (MHC) molecules that present peptide antigens to T-cell receptors:
In the CNS, microglial cells express both HLA class I and class II molecules, enabling them to present antigens and interact with T cells. This capacity is enhanced in neurodegenerative conditions where inflammatory responses are activated.
Prior to the 2026 study, HLA associations with PSP included:
The Forrest et al. (2026) study expanded on these findings by investigating functional consequences rather than merely genetic associations.
Forrest et al. (2026) analyzed 32 PSP cases with known HLA haplotypes [1:1]:
The study applied machine learning to identify patterns invisible to conventional statistical methods:
The machine learning models revealed that specific ratios of neuroinflammatory markers reliably distinguished HLA haplotypes, exceeding the performance of individual biomarkers alone.
Between HLA-defined groups, neuropathological analysis revealed significant regional differences:
| Brain Region | Key Finding |
|---|---|
| Basal ganglia | Differential microglia load by HLA haplotype |
| Subthalamic nucleus | Varying p-Tau burden across HLA groups |
| Frontal cortex | Different cytotoxic T cell density |
| Brainstem | Variable astrocyte activation patterns |
Machine learning identified that the ratio of neuroinflammatory markers (not absolute values) most reliably distinguished HLA haplotypes. Key discriminators included:
These ratios suggest that HLA haplotypes influence the type and balance of immune response rather than simply its intensity.
Perhaps the most clinically relevant finding: symptom progression sequences differed by HLA haplotype. This indicates that:
The study supports the notion that PSP pathology may be associated with various aetiological-pathogenic events, including rareyly occurring targetable autoimmune mechanisms. This expands the conceptual framework of PSP beyond pure tauopathy to include an immune-mediated component in some cases.
The HLA-dependent neuroinflammatory profiles have direct implications for disease-modifying therapy trials:
The findings support exploration of immunomodulatory strategies in PSP:
| Approach | Rationale | Target Population |
|---|---|---|
| HLA-targeted immunotherapy | Antigen presentation dysregulation | HLA-risk haplotype carriers |
| T cell modulation | Cytotoxic T cell involvement | High CD8+ density groups |
| B cell-targeted therapy | B cell participation | DRB115:01-DQB106.02 carriers |
| Microglial modulation | HLA-dependent microglial profiles | Specific ratio-based groups |
HLA associations in PSP differ from those in Alzheimer's disease:
| Feature | PSP (Forrest 2026) | AD |
|---|---|---|
| HLA association strength | Moderate, specific | Strong, widespread |
| Primary immune cell type | Mixed (T, B, microglia) | Microglia-dominated |
| TREM2 interaction | Weaker | Stronger |
| HLA-risk haplotype effect | Phenotype-modifying | Risk-modifying |
| Autoimmune component | Present in subset | Minimal |
HLA Haplotype Variants
|
v
Differential Antigen Presentation Capacity
|
+---> CD8+ T Cell Activation ---> Cytotoxic T Cell Infiltration
|
+---> Microglial HLA Expression ---> Altered Phagocytosis
|
+---> Cytokine Profile ---> Neuroinflammatory Balance
|
v
Distinct Neuroinflammatory Phenotype
|
v
Symptom Progression Sequence Differences
|
v
Clinical Heterogeneity within PSP
Several non-mutually exclusive mechanisms may explain the HLA-PSP relationship:
The 2026 study opens several research avenues:
The findings complement single-cell microglial profiling studies [3]:
Work is needed to identify peripheral correlates of HLA-dependent neuroinflammatory profiles:
Forrest SL, Zaheer SS, Kim A, Tanaka H, Chasiotis H, Li J, Fox SH, Wang J, Tartaglia MC, Lang AE, Kovacs GG. Distinct neuroinflammatory profiles in progressive supranuclear palsy associated with HLA haplotypes. Brain Communications. 2026. ↩︎ ↩︎
Zouridis JM, et al. HLA haplotypes in progressive supranuclear palsy. Neurology. 2017. ↩︎
Chen J, et al. Single-cell microglial profiling in PSP. Nature. 2024. ↩︎