Baseline Inflammatory Profiles in Moderate-to-Severe Depression and Differential Response to Intermittent Theta-Burst Stimulation

This exploratory study demonstrates that baseline inflammatory protein signatures, particularly those involving cytokine and chemokine signaling pathways, can stratify patients with moderate-to-severe depression into distinct subgroups with differential responsiveness to intermittent theta-burst stimulation (iTBS), suggesting a potential biomarker-driven approach for personalized psychiatric treatment.

The Gist

The Big Picture: Why Do Some Depression Treatments Work and Others Don't?

Imagine you have a broken car. You take it to two different mechanics. Mechanic A fixes it perfectly, but Mechanic B can't get it running at all. You might think, "Well, the car is the same, so the fix should be the same." But what if Mechanic A was fixing a car with a flat tire, while Mechanic B was trying to fix a car with a flooded engine? They are both "broken cars," but the type of breakage is totally different.

This is exactly the problem with Major Depressive Disorder (MDD). Doctors often treat all depression the same way, but for many people, the treatment doesn't work. This study suggests that depression isn't just one thing; it's a mix of different "flavors," and one of those flavors is inflammation (your body's immune system being on high alert).

The researchers wanted to see if they could predict who would get better from a specific brain treatment called iTBS (a type of magnetic stimulation that "reboots" brain circuits) by looking at the patient's immune system before they started.

---

The Experiment: Sorting Patients by Their "Immune Fingerprint"

1. The Treatment (iTBS)
Think of the brain as a garden that has become overgrown and tangled. iTBS is like a high-tech gardener using a precise tool to trim the weeds and help the flowers grow back. It's a non-invasive treatment that uses magnetic waves to stimulate the brain.

2. The Blood Test
Before the treatment started, the researchers took blood samples from 54 patients. Instead of just looking at one or two things, they used a super-sophisticated scanner (called the Olink panel) to check 92 different immune proteins at once.

• Analogy: Imagine checking the weather. Instead of just asking "Is it raining?", they checked the temperature, humidity, wind speed, barometric pressure, and UV index all at once to get a full picture of the "climate" inside the patient's body.

3. The Sorting (Clustering)
Using a computer algorithm, they sorted the patients into two groups based on their immune "climate":

• Group 1 (The "Stormy" Group): These patients had high levels of inflammatory proteins. Their immune systems were like a fire alarm that wouldn't stop screaming.
• Group 2 (The "Calm" Group): These patients had lower levels of inflammation. Their immune systems were more like a quiet hum.

---

The Results: Who Got Better?

The results were striking and revealed a clear pattern:

• Group 2 (Calm Immune System): Most of these patients got significantly better after the iTBS treatment. The "gardener" was able to fix the garden because the soil wasn't too acidic or toxic.
• Group 1 (Stormy Immune System): Most of these patients did not get better. The treatment didn't work well.

Why?
The researchers believe that when the immune system is in "overdrive" (high inflammation), it creates a toxic environment in the brain. It's like trying to plant seeds in soil that is on fire. The brain's ability to change and heal (called neuroplasticity) gets blocked by the inflammation. The iTBS treatment tries to help the brain grow new connections, but the inflammation acts like a shield, stopping the treatment from working.

Key Finding:
The two groups looked exactly the same on paper. They had the same age, same gender, same severity of sadness, and were taking the same medications. The only thing that predicted who would get better was their invisible immune profile.

---

The "Smoking Guns": What Proteins Were Different?

The study found 10 specific proteins that were much higher in the "Stormy" group (the ones who didn't get better).

• The Culprits: These included proteins like IL-8, MCP-4, and MMP-1.
• The Analogy: Think of these proteins as "construction workers" who are supposed to help repair the brain. But in these patients, they are acting like a chaotic mob. They are tearing down the walls (damaging brain connections) and blocking the road (preventing new growth) instead of helping.

The study also looked at the "pathways" these proteins use. It turns out they are part of a communication network called Cytokine Signaling. In the "Stormy" group, this network is shouting too loudly, confusing the brain's repair crew.

---

Why Does This Matter? (The Future of Psychiatry)

This study is a big step toward Precision Psychiatry.

• Old Way: "Here is a treatment. Try it. If it doesn't work, try another one." (This is like guessing which mechanic to call).
• New Way: "Let's check your immune profile first. If you have high inflammation, we might need to treat the inflammation before we try the brain stimulation, or we might choose a different treatment entirely."

The Takeaway:
Depression is complex. Sometimes, the reason a treatment fails isn't because the treatment is bad, but because the patient's body is fighting a hidden battle (inflammation) that blocks the cure. By identifying these "immune fingerprints" early, doctors might be able to personalize treatment, saving patients from years of trial and error.

In short: If your brain is a garden, and the soil is on fire (inflammation), you can't just plant seeds (treatment) and expect flowers. You have to put out the fire first. This study helps us identify who has the fire so we can put it out before we try to grow the garden.

Technical Summary

1. Problem Statement

Major Depressive Disorder (MDD) is biologically heterogeneous, with a recognized "inflammatory subtype" characterized by elevated pro-inflammatory cytokines (e.g., IL-6, TNF-α) and associated with poorer responses to antidepressants. While repetitive transcranial magnetic stimulation (rTMS), specifically intermittent theta-burst stimulation (iTBS), is an effective non-invasive treatment, response rates remain highly variable. Current literature lacks a comprehensive, systems-level understanding of how baseline immune profiles influence neuromodulation outcomes. Most existing studies focus on single cytokines, yielding inconsistent results. This study addresses the gap by investigating whether baseline inflammatory protein signatures can stratify patients with moderate-to-severe depression into distinct subgroups that differ in their response to iTBS.

2. Methodology

Study Design and Participants:

• Cohort: 54 adults (aged 18–60) with moderate-to-severe MDD (baseline MADRS ≥ 20).
• Intervention: Participants underwent a six-week, quadruple-blinded, randomized, sham-controlled crossover trial involving active iTBS and sham stimulation.
• Targeting: Stimulation was applied to the left dorsolateral prefrontal cortex (DLPFC), either at the standard F3 location or an individually targeted site via fMRI (both groups combined for this analysis).
• Exclusion: Patients with acute systemic inflammation (CRP > 10 mg/L or WBC > 11,000 cells/μL) were excluded to isolate chronic low-grade inflammation.

Proteomic Profiling:

• Platform: Olink® Target 96 Inflammation panel.
• Metrics: Quantification of 92 inflammation-related proteins using Proximity Extension Assay (PEA) technology.
• Data Handling: Data were presented as Normalized Protein Expression (NPX) values. No imputation was performed for values below the Limit of Detection (LOD) to preserve natural biological variance; all 92 proteins were retained.

Statistical Analysis:

• Clustering: K-means clustering (Hartigan–Wong algorithm) was applied to standardized NPX values to identify patient subgroups. Optimal cluster number ($k=2$) was determined via Elbow method, Gap Statistics, and Silhouette width. Stability was confirmed via bootstrap resampling (Jaccard indices > 0.95).
• Differential Expression: Volcano plots identified differentially expressed proteins (DEPs) between clusters using Welch's t-tests. Significance thresholds were set at FDR < 0.05 and |ΔNPX| ≥ 1 (approx. 2-fold change).
• Pathway Analysis: KEGG and Reactome databases were used for enrichment analysis of the DEPs.
• Confounding Control: Sensitivity analyses (Firth penalized logistic regression) controlled for medication class, medication load, and other clinical variables.

3. Key Contributions

• Data-Driven Stratification: Moves beyond symptom-based classification to identify biologically distinct "immune biotypes" in MDD using high-throughput proteomics.
• Systems-Level Approach: Instead of analyzing single markers, the study characterizes complex, interacting immune pathways associated with treatment resistance.
• Neuromodulation Specificity: Provides the first evidence linking specific baseline inflammatory signatures to differential outcomes in iTBS, extending the known relationship between inflammation and pharmacotherapy resistance to neuromodulation.

4. Key Results

Cluster Characteristics:

• Cluster 1 (n=28): Characterized by lower iTBS response rates (28.6% responders).
• Cluster 2 (n=26): Characterized by higher iTBS response rates (57.7% responders).
• Demographics: No significant differences existed between clusters regarding age, sex, BMI, baseline MADRS scores, medication status, or treatment resistance history.
• Outcome: While the mean reduction in MADRS scores was similar between groups, Cluster 2 had a significantly higher proportion of categorical responders (≥50% symptom reduction) ($p = 0.031$).

Differentially Expressed Proteins:
Ten proteins met the strict significance criteria (FDR < 0.05 and |ΔNPX| ≥ 1), all showing higher expression in the low-response Cluster 1:

1. MCP-4 (CCL13)
2. CXCL6
3. IL-7
4. TNFSF14 (LIGHT)
5. CXCL1
6. OSM (Oncostatin M)
7. CXCL5
8. IL-8
9. CXCL11
10. MMP-1 (Matrix Metalloproteinase-1)

Pathway Enrichment:
Enrichment analysis revealed that the low-response profile was driven by:

• Cytokine–cytokine receptor interactions (KEGG FDR = $9.94 \times 10^{-12}$).
• Chemokine signaling and chemokine receptor binding (Reactome FDR = $9.11 \times 10^{-11}$).
• GPCR-mediated signaling, IL-17 signaling, NF-κB signaling, and JAK-STAT signaling.

Effect Sizes:
The differences were robust, with large effect sizes (e.g., Cohen's $d$ for MCP-4 = 3.18; IL-7 = 2.40).

5. Significance and Implications

Biological Mechanism:
The findings suggest that a pro-inflammatory baseline state, specifically involving chemokine-driven leukocyte recruitment (IL-8, CXCLs) and extracellular matrix remodeling (MMP-1), may impair the neuroplasticity required for iTBS to be effective. High levels of these markers may indicate a "locked" neuroimmune state that resists synaptic strengthening.

Clinical Translation:

• Precision Psychiatry: Baseline inflammatory profiling could serve as a biomarker to predict who will benefit from iTBS versus those who might require alternative or augmented treatments (e.g., anti-inflammatory augmentation).
• Treatment Resistance: Identifies a specific biological mechanism for treatment resistance in neuromodulation, distinct from general symptom severity.

Limitations:

• Exploratory Nature: The study is hypothesis-generating; causal inference is limited by the small sample size ($N=54$) and lack of a non-iTBS control arm for the proteomic analysis.
• Peripheral vs. Central: Proteins were measured in plasma; while peripheral inflammation correlates with neuroinflammation, direct central mechanisms were not measured.
• Generalizability: Findings are specific to moderate-to-severe depression and may not apply to mild cases.

Conclusion:
This study demonstrates that baseline inflammatory proteomic profiles can stratify patients with MDD into subgroups with distinct iTBS outcomes. Patients with elevated chemokine and cytokine signaling (Cluster 1) show poorer response, suggesting that immune dysregulation is a critical biological modifier of neuromodulatory efficacy. Future research should validate these markers in larger cohorts and explore anti-inflammatory strategies for this specific subgroup.