Can THC Reduce Arthritis-Cell Inflammation While Increasing Other Immune Responses? What 14 Synovial-Tissue Donors Show

Short answer: In a laboratory study using synovial fibroblasts isolated from 14 women undergoing knee replacement for rheumatoid arthritis, Δ9-tetrahydrocannabinol (THC) produced concentration- and cell-context-dependent effects. In tumor-necrosis-factor-primed fibroblasts, micromolar THC increased intracellular calcium and uptake of a fluorescent probe through a pathway responsive to TRPA1 antagonists. At the highest concentration tested, 25 µM, it lowered matrix metalloproteinase-3 (MMP-3) release but also caused pronounced loss of viable cells. In cultures containing immune cells, some tumor necrosis factor (TNF) signals decreased while selected immunoglobulin G (IgG) responses increased. These results do not establish that cannabis treats rheumatoid arthritis, suppresses inflammation in a patient, produces a predictable human blood THC concentration, or proves an individual’s impairment.

The primary evidence is Lowin, Kok, Smutny and Pongratz (2022), Impact of Δ9-Tetrahydrocannabinol on Rheumatoid Arthritis Synovial Fibroblasts Alone and in Co-Culture with Peripheral Blood Mononuclear Cells, Biomedicines 10(5):1118, doi:10.3390/biomedicines10051118. The distinction that matters is between an effect on cultured cells and a clinically measured response in a living person.

What did the researchers actually study?

The study enrolled 14 women with established rheumatoid arthritis, mean age 68 ± 10 years, who were already undergoing elective knee-joint replacement. Their synovial tissue supplied rheumatoid arthritis synovial fibroblasts (RASF). This was tissue donated during clinical surgery, not a trial in which those 14 women were prescribed or administered THC. Medication histories were heterogeneous, including glucocorticoids, methotrexate, biologics and other treatments. Those clinical characteristics describe the tissue source; they do not transform the cell experiment into a randomized clinical treatment trial.

Investigators exposed isolated fibroblasts to laboratory THC solutions spanning 0.1–25 micromolar (µM), sometimes after priming the cells for 72 hours with TNF at 10 ng/mL. They also isolated peripheral blood mononuclear cells (PBMC) from healthy donors and maintained those cells alone or together with rheumatoid fibroblasts. The PBMC experiments evaluated 1 and 10 µM THC over seven-day stimulation periods. The number of cultured wells, technical readings and sampled cells is not the number of independent rheumatoid arthritis patients; the figure-level sample counts differ, including approximately 8–11 experimental observations for selected calcium/probe experiments, five for some cytokine comparisons and three for immunoglobulin assays.

The design therefore offers a mechanistic comparison between types of living human-derived cells in culture. It does not measure symptomatic joint improvement, drug pharmacokinetics, long-term immune health, disability, pain scores, intoxication or driving. Laboratory concentrations are input conditions, not measured human doses or forensic blood results.

Which specimens and measurements were real?

The biological starting material was knee synovial tissue, from which fibroblasts were cultured. The additional immune-cell material was peripheral blood-derived mononuclear cells from healthy donors, isolated before being placed in culture. Once isolated, those cells were studied in wells rather than as circulating blood. Researchers used flow cytometry for cannabinoid-receptor expression, Cal-520 fluorescence for intracellular calcium, PoPo3 fluorescence as a probe-uptake surrogate, ELISA for cytokines and immunoglobulins in conditioned medium, and a cell-viability assay.

No study participant received a measured systemic THC dose. No venous whole-blood THC, serum THC, plasma THC, oral-fluid THC, urinary THC metabolite, exhaled THC or vitreous THC concentration was reported as a pharmacokinetic endpoint. A solution labeled 10 µM THC in a cell well cannot be substituted for 10 µM in an individual’s bloodstream. It is equally invalid to infer a blood-alcohol-equivalent concentration or a legally meaningful THC cutoff from this culture experiment.

This distinction is consistent with the underlying specimen principle explained in the site’s guide to biological matrices in forensic toxicology: an analytical or biological result remains tied to the material and method actually examined.

Why did tumor-necrosis-factor priming change the THC response?

Untreated fibroblasts showed little calcium response when exposed to 0.1–25 µM THC. After 72 hours of TNF priming, however, THC at 5–25 µM increased intracellular calcium fluorescence by up to approximately 200% of the comparator response; the authors report p < .001 for those concentrations. This contrast makes pretreatment context essential. Saying simply that “THC raises calcium in arthritis cells” removes the experimental condition responsible for the strong result.

Flow cytometry identified mainly intracellular CB1 signal and cell-surface CB2; TNF increased surface CB2 with a reported boundary p = .05. Calcium mobilization did not disappear with the CB1 inverse agonist rimonabant. Two lipophilic antagonists of the transient receptor potential ankyrin 1 channel (TRPA1), HC030031 and A967079, markedly reduced the THC-induced response. A967079 reduced responses across 5–25 µM THC at p < .001 in the reported comparisons.

These inhibitor patterns support a role for TRPA1-linked signaling; they do not independently prove that every affected molecule traveled exclusively through TRPA1 or that an identical mechanism dominates inside a patient’s inflamed joint. Pharmacological blockers can have off-target or concentration-dependent actions, and receptor expression alone is not a validated treatment outcome.

Did calcium still change without extracellular calcium?

Yes, in the model. Investigators replaced calcium-containing Hanks’ balanced salt solution with calcium-free phosphate-buffered saline for a separate set of experiments. The response pattern persisted in TNF-primed fibroblasts, consistent with an intracellular calcium-source contribution rather than simple influx from the surrounding solution. Under those conditions the TRPA1 antagonist A967079 effectively abrogated the observed THC calcium response in the reported series.

This is an important mechanistic qualifier, not a clinical dose-response curve. The study measured changing fluorescence from a dye in living cultured cells over approximately 90 minutes. It did not directly image a patient’s joints during cannabis use or quantify neurobehavioral intoxication. Similarly, a calcium trace is not an assessment of mood, motor control, inflammation severity or drug-induced disability.

Did THC make arthritis fibroblasts take up more drug?

The result was an increased signal from PoPo3, a fluorescent surrogate probe, not an assay of an antirheumatic drug actually delivered into a patient’s joint. TNF-primed cultures demonstrated increased probe uptake at THC concentrations of 1–25 µM, with reported p < .001 in the relevant comparison. TRPA1 blockers inhibited much of that uptake. The organic-cation transport inhibitor decynium-22 also reduced uptake, including with THC present, while some CB2-related pharmacology changed the signal under specific conditions.

The authors investigated this phenomenon because the ability of cells to internalize a tracer could inform mechanisms of transport. It would nevertheless be incorrect to conclude that THC has been proven to enhance delivery of methotrexate or another disease-modifying drug in human arthritis. PoPo3 fluorescence is an exploratory stand-in for transport behavior. Actual antirheumatic compound identity, intracellular drug concentrations, therapeutic efficacy and adverse reactions were not measured in the reported uptake experiment.

Evidence table 1: the fibroblast and probe findings

Did THC suppress the inflammatory proteins IL-6, IL-8 and MMP-3?

Not as a uniform class-wide effect. In the TNF-primed fibroblast cultures, the researchers did not find significant suppression of IL-6 or IL-8 with THC alone across the tested doses. By contrast, MMP-3 release was strongly reduced at 25 µM THC (p < .001), including certain antagonist co-exposure conditions. IL-8 was significantly reduced at that same high THC concentration when combined with A967079 (p < .001) or rimonabant (p = .005), rather than demonstrating broad suppression by THC itself.

Figure 4 of the original paper expresses IL-6, IL-8 and MMP-3 protein in culture-medium concentrations of ng/mL. Those values are not parent THC concentrations and are not interchangeable with concentrations obtained in a person’s blood. The distinction matters because an immune mediator falling in a culture could reflect changing cell signaling, cell survival or both. The observed anti-inflammatory-sounding output must be evaluated with the accompanying viability findings rather than treated in isolation.

Why does the cell-death result change the interpretation?

At intermediate THC concentrations, the cell-viability assay showed a small increase in viable-cell signal—for example, at 5 µM the authors report p = .022. At 25 µM, however, there was a dramatic decrease in viable-cell number (p < .001). A culture with many fewer surviving cells can release less MMP-3 or other proteins for reasons very different from a selective therapeutic reduction in inflammatory signaling.

This is the central competing explanation for any claim that the highest THC concentration “treated inflammation.” The experiment cannot separate a beneficial signaling change from a component attributable to cytotoxicity simply by observing less protein in a conditioned medium sample. Without parallel cell-normalized secretion, viability-adjusted measures and disease endpoints, it is unsound to equate low MMP-3 with less arthritis in a person.

The authors themselves discuss impaired viability at high THC exposure as a possible contributor to reduced secreted mediators. A scientifically balanced reading preserves both findings: the biochemical endpoint changed, but the cellular system was also damaged at that concentration.

What changed when healthy-donor immune cells were added?

The study added healthy-donor PBMC to rheumatoid synovial fibroblasts to examine cell-cell interaction. The co-cultures used approximately 5,000 fibroblasts and 250,000 PBMC per well and assessed cytokines and immunoglobulins after seven days with selected immune stimuli. The different proportions of cells influence which cell type contributes most of a measured protein. They are not numbers of independent human donors or evidence that hundreds of thousands of people were sampled.

With 1 µM THC in co-culture, TNF output declined under CpG stimulation (p = .027) and interferon-γ stimulation (p = .010) in the authors’ reported comparisons. This response was not replicated simply as a broad TNF suppression across every isolated-cell condition: the same experimental context mattered. Some IL-6 and IL-10 comparisons showed no consistent THC-related effect, while the original figure’s annotations should be interpreted with their individual stimulus-specific tests rather than treated as an omnibus null.

Healthy-donor PBMC and rheumatoid-patient fibroblasts therefore cannot be collapsed into one participant group. The cell mixture is deliberately reconstructed in vitro, under controlled stimulation, rather than a clinical record of a patient receiving a THC formulation.

Could THC increase one immune response while decreasing another?

Yes. In isolated PBMC cultures, 10 µM THC lowered IL-10 under interferon-γ stimulation (p = .011) and CpG stimulation (p = .03). In an unstimulated PBMC setting, 1 µM THC increased IgG production (p = .026). By contrast, IgG increases in CpG-stimulated co-cultures did not reach the conventional significance threshold at either 1 µM (p = .077) or 10 µM (p = .085).

The direction of these findings is not logically contradictory: the experiments used different immune-cell compositions, activating stimuli and outcome proteins. TNF, IL-10 and IgG are not interchangeable measures of one universal inflammatory state. Suppressing a proinflammatory mediator under one condition does not prove global immune suppression; raising IgG in a culture does not prove improved host defense, autoantibody activity, arthritis progression or protection from infection.

In fact, this mixture of findings is a reason to resist a one-word label such as “anti-inflammatory.” For forensic assessment, immune signaling may be relevant to mechanism or toxicity hypotheses, but it does not demonstrate what occurred in a particular person’s immune system after cannabis exposure.

Evidence table 2: immune co-culture outcomes and true null results

Does this show THC works through CB1, CB2 or TRPA1?

It supports more than one possible signaling route. TRPA1 antagonist sensitivity is strong evidence for involvement of that channel in calcium mobilization and fluorescent probe uptake in TNF-primed fibroblasts. CB2 appeared on fibroblast surfaces and its expression responded to TNF; antagonist experiments suggested some context-specific contribution. PBMC have a different repertoire of receptors and activation pathways. The study did not perform an exhaustive receptor knockout program or prove that a single receptor exclusively mediates every cytokine, IgG and viability endpoint.

Concentrations also matter: the study included micromolar THC conditions that were selected for laboratory pathway interrogation. A patient taking prescribed oral dronabinol or using a cannabis product cannot be assumed to expose an arthritic joint to the same unbound concentration over the same duration. No human THC absorption, distribution, binding, metabolism or tissue penetration was measured to bridge that gap. Mechanistic plausibility is not a validated effective dose, safe threshold or therapeutic efficacy statement.

Could PoPo3 uptake support a drug-interaction argument?

Only as a hypothesis-generating observation. The authors suggest that changing membrane or transport behavior could influence cellular uptake of a co-administered agent, but the tested probe was not a standard disease-modifying antirheumatic therapy. The cell population was experimentally TNF-primed. Follow-up work would need to quantify each named medication, its transport kinetics, intracellular concentration, functional benefit and toxicity at exposure levels actually achievable in people.

More generally, an interaction allegation in a legal matter should begin with the medicines involved, dose history, specimen collection and validated methods. The site’s specimen-collection guide discusses why the collection context and material cannot be separated from the resulting analytical claim. A cell fluorescence response alone does not identify the concentration of an interacting drug in blood or tissue.

How should this paper be used in toxicology testimony or litigation?

It can help explain that cell-culture toxicology is not equivalent to an individual’s exposure or functional state. The independent patient tissue source, healthy donor immune cells, number of experimental wells, laboratory THC concentrations, experimental stimuli and outcome instruments should each be stated accurately. The difference between a significant conditioned-medium TNF result and a nonsignificant IgG comparison is meaningful. So is the finding that high-concentration MMP-3 suppression coexisted with striking cell loss.

It cannot prove that a patient was helped or injured by cannabis, that rheumatoid inflammation improved, that a particular plasma or whole-blood THC value would produce these responses, that prescribed medicine was absorbed more efficiently, or that a person was impaired at a legally relevant time. Those propositions require different forms of evidence, potentially including clinical endpoints, observed performance, measured drug concentrations in an appropriate matrix, longitudinal history and alternative causes. A confirmed chemical result would likewise remain separate from biological effect; the confirmatory-testing guide explains the analytical distinction.

What are the main limitations and plausible competing interpretations?

The fibroblast source consisted of 14 older female surgical patients with long-standing rheumatoid arthritis and varied prior medication exposure. That limits demographic generalization. Healthy-donor PBMC supplied a different component of the co-culture. Technical wells and repeated measurements do not add independent people. Many treatment conditions, antagonists and proteins were compared, making individual p values dependent on the study’s analysis choices and multiplicity handling. Figure-level sample sizes differ, and detailed clinical outcomes were not collected.

THC was applied directly to culture medium; no living patient received an observed systemic THC exposure. Fluorescence represented calcium dye or tracer uptake, not a directly measured drug concentration within a person. High concentrations changed cell viability. Immunologic endpoints were heterogeneous, including both increases and decreases depending on the stimulus and cell composition. These limitations favor a mechanistic, exploratory interpretation rather than a demonstrated intervention, diagnosis, impairment rule or patient-specific causation opinion.

Frequently asked forensic and scientific questions

1. Did 14 rheumatoid arthritis patients receive THC treatment?

No. Synovial tissue was obtained from 14 women undergoing knee replacement. THC was applied afterward to cells maintained in laboratory culture. The paper was not a clinical trial of treating those women with THC.

2. Does a lower MMP-3 result prove that THC safely reduced inflammation?

No. The pronounced MMP-3 reduction occurred at 25 µM, where viable fibroblast numbers also fell sharply. Cell loss offers a competing explanation and there were no patient-level arthritis outcomes.

3. Was IL-6 consistently suppressed by THC in fibroblasts?

No. The original investigators did not report a significant uniform THC-alone suppression of fibroblast IL-6 under their TNF-primed conditions.

4. Can this study be used to predict a positive blood or urine THC test?

No. The concentrations were prepared in cell-culture media, and the study did not collect a pharmacokinetic series of parent THC or metabolites in human whole blood, plasma, serum or urine after an administered dose.

5. Does increased PoPo3 fluorescence prove another medication will enter the joint?

No. PoPo3 was a surrogate tracer under specific culture conditions. No named antirheumatic medicine was directly quantified for enhanced delivery or therapeutic benefit.

6. Did the PBMC experiments establish overall immune suppression?

No. Some co-culture TNF and monoculture IL-10 signals fell, while isolated-PBMC IgG increased under another condition. Several other contrasts, including stimulated co-culture IgG, were nonsignificant.

7. Does any result establish impairment, driving risk or causation in a case?

No. There was no intoxicated-driving test, neurobehavioral impairment assessment, individual blood THC threshold or case-specific outcome. A cellular mechanism cannot replace direct evidence about the person and event at issue.

Primary-paper conclusion

Lowin and colleagues showed that THC produced conditional and sometimes opposing laboratory effects in rheumatoid synovial fibroblasts and peripheral immune cells. TRPA1-sensitive calcium and tracer uptake were prominent in TNF-primed fibroblasts. At high THC concentration MMP-3 fell alongside marked cytotoxicity, while mixed-cell immune experiments revealed both reductions and increases in different endpoints and several genuine null comparisons. The appropriate conclusion is mechanistic and bounded, not a clinical claim that THC reliably treats arthritis or establishes forensic impairment. Review the complete original study: Lowin et al., 2022, DOI 10.3390/biomedicines10051118.