Compounds with a graded row

Ordered by rung, strongest first. The rung rates the study design; the chip beside it says what the study measured. Both are needed.

1. Adiponectin

   animal in vivo Direct outcome Not approved for this condition

   Who was studied

   DBA/1J mice, 8-12 weeks old, sex not reported in the source, n=7 per group for arthritis scoring and n=3 for histology and micro-CT, collagen-induced arthritis; globular adiponectin 10 mcg in 10 mcL of PBS injected into the knee joints on days 17, 20 and 23 after the first immunisation

   This row records a harm signal, and a contested one. Intra-articular adiponectin brought arthritis on earlier and made it worse - more synovial hyperplasia, more bone erosion, lower trabecular bone mineral density on micro-CT, a roughly 1.5-fold expansion of joint Th17 cells and a tenfold rise in IL-17 mRNA by day 45. The isoform given was the globular domain, not full-length adiponectin, and the authors state that a separate group reported the opposite direction in the same model and attribute the disagreement to that isoform difference, so the record here belongs to globular adiponectin specifically. No human study administering adiponectin in rheumatoid arthritis is on file; the human literature measures serum and synovial adiponectin as a marker.

   Source PubMed Central 1 primary source read for this row

2. Angiotensin (1-7)

   animal in vivo Direct outcome Not approved for this condition

   Who was studied

   male DBA/1 mice, 7 weeks old, n=10 per group across 6 groups, collagen-induced arthritis; angiotensin-(1-7) 2.0 mg/kg intraperitoneally once daily for 7 days from day 28, started after arthritis onset

   Clinical arthritis score, paw thickness and ankle joint histology were all measured directly and all improved, alongside lower TNF-alpha, IL-1beta, IL-6 and CRP and reduced RANKL and MMP-3 in joint tissue. The molecule given was plain angiotensin-(1-7); a separate rat adjuvant-arthritis study used a bone-targeting conjugate, which is a different construct, and its result does not transfer to the bare peptide. Human trials of angiotensin-(1-7) do exist - a seamless phase 1-2 randomised trial infused it into 107 COVID-19 intensive care patients and measured oxygen-free days - but none measured an arthritis outcome, and no human arthritis trial is on file.

   Source PubMed Central 2 primary sources read for this row

3. Bactericidal/Permeability-Increasing Protein (BPI)

   animal in vivo Direct outcome Not approved for this condition

   Who was studied

   male C57Bl/6 mice, 9-10 weeks old, n=12 per condition, collagen-induced arthritis; recombinant BPI at 50 mcg/mL in 200 mcL intraperitoneally twice weekly for 2 months, started only once a mouse had reached arthritis score 2 (days 28-33 after immunisation)

   Histological scores for pannus formation and inflammation of the joint fell two- to fourfold and serum TNF-alpha was about fifteenfold lower than in the vehicle group. The molecule given was recombinant full-length BPI, not the 21-kDa amino-terminal fragment rBPI21 that went into the paediatric meningococcal sepsis trial, so the two records should not be pooled. The authors state plainly that BPI has never been evaluated clinically in arthritis or a comparable condition, and no registered arthritis trial of BPI is on file.

   Source PubMed Central 1 primary source read for this row

4. Substance P

   animal in vivo Direct outcome Not approved for this condition

   Who was studied

   rats with adjuvant-induced arthritis, substance P infused into the knee - the joint that normally develops the milder disease; strain, sex, group sizes and infusion duration are not reported in the record that could be read

   This row records a harm signal rather than a benefit. Infusing substance P into the knee increased the severity of adjuvant-induced arthritis, while injecting a substance P receptor antagonist did not change it, and the joints that developed the more severe disease were the more densely innervated by substance P-containing primary afferents. Only the indexed PubMed record could be read rather than the full Science paper, so the animal detail above is what the record states; no study administering substance P to humans with rheumatoid arthritis exists, and the human work measures it in synovial fluid instead.

   Source PubMed 1 primary source read for this row

5. VIP

   animal in vivo Direct outcome Not approved for this condition

   Who was studied

   male DBA/1J mice, 6-10 weeks old, n=10 per group, collagen-induced arthritis; VIP 1 nmol per animal intraperitoneally on alternate days from day 25 to day 35, started after clinical arthritis had already appeared

   The joints themselves were scored, not a marker: the clinical arthritis score fell from 5.08 to 1.56 and the histological bone erosion score from 2.13 to 0.25, with the joint RANKL/OPG ratio falling roughly sevenfold, and the earlier report in Nature Medicine (Delgado 2001, PMID 11329057) found the same direction in the same model. Every human VIP finding in rheumatoid arthritis on file measures the peptide rather than administers it - in a 91-patient early-arthritis cohort followed to five years, low baseline serum VIP marked a worse disease course, which is observational biology about a molecule the body already makes. A ClinicalTrials.gov query for VIP or aviptadil against arthritis returns no study in which the peptide is given to anyone with arthritis - only an observational blood-sampling study that lists rheumatoid arthritis among its comparison groups.

   Source PubMed Central 3 primary sources read for this row

Recorded absences

6 compounds are listed here without a rung. Each row exists to mark a distinction a reader would otherwise draw wrongly, and each says what is missing rather than leaving the compound off the page.

1. CRH (Corticotropin-Releasing Hormone)

   No rung — nothing on file to grade Not approved for this condition

   Nothing administered located. The nearest human finding is for the synthetic analogue corticorelin acetate against a steroid burden rather than against arthritis: a randomised, double-blind trial in 200 patients with peritumoral brain oedema on chronic dexamethasone found 1 mg twice daily subcutaneously allowed a greater maximum steroid reduction than placebo (62.7% versus 51.4%, P<0.001) with less myopathy, while the primary responder endpoint missed significance (57.0% versus 46.0%, P=0.12) (https://pubmed.ncbi.nlm.nih.gov/23382470/). Peritumoral cerebral oedema is not a condition in this corpus, so the trial is recorded here rather than graded.

2. Hyaluronic Acid Binding Peptide

   No rung — nothing on file to grade Not approved for this condition

   No study has used a hyaluronic-acid-binding peptide against rheumatoid arthritis. The joint work here is entirely post-traumatic osteoarthritis in mice.

3. KPV

   No rung — nothing on file to grade Not approved for this condition

   This row records an absence. ClinicalTrials.gov returns no registered study of KPV for any condition in any country, and the published in vivo work on the tripeptide sits in other diseases: crystal-induced and IL-1beta-induced peritonitis in mice, where the outcome measured was neutrophil accumulation in the peritoneal cavity, and a mouse colitis model entered through the PepT1 transporter. No study on file measured a joint, a swollen-joint count, an ACR response, a DAS28 score or joint histology in any arthritis model, in any species. The anti-inflammatory claim attached to this compound in the market rests on the peritonitis and colitis work and on its parent hormone alpha-MSH, not on a rheumatoid arthritis result.

4. LL-37 (Cathelicidin)

   No rung — nothing on file to grade Not approved for this condition

   This row records an absence, and the reason is a fragment. The one arthritis intervention study on file gave IG-19, an internal segment of LL-37, to mice with collagen-induced arthritis and reported lower disease severity, less cellular infiltration in the joints and lower serum antibodies against type II collagen - but IG-19 is not LL-37, and a bovine cathelicidin-derived peptide tested alongside it did not reproduce the effect, which is exactly why the evidence stays with the fragment. Where LL-37 itself appears in the rheumatoid arthritis literature it is measured rather than given: expression in synovial membrane, and serum autoantibodies against it. The one randomised human trial that did administer LL-37 applied it topically to venous leg ulcers and measured wound healing.

5. PEA

   No rung — nothing on file to grade Not approved for this condition

   This row records an absence, and it is the informative kind: palmitoylethanolamide has a large randomised human literature and none of it is here. A 2024 systematic review catalogued 47 randomised controlled trials across neuropsychiatric, neurological, somatic and visceral conditions and none was in rheumatoid arthritis. The inflammatory-marker results that review collects belong to other populations - in an 8-week quadruple-blind trial in 66 adults with diabetic peripheral neuropathy, serum IL-6 fell relative to placebo (p=0.04) while CRP did not (p=0.261) - so they are evidence about those conditions, not this one. PEA is also a fatty-acid amide rather than a peptide, despite this record carrying the peptide flag.

6. TB-500

   No rung — nothing on file to grade Not approved for this condition

   This row records an absence. A ClinicalTrials.gov query for thymosin beta-4 and TB-500 returns nineteen registered studies - dry eye, neurotrophic keratopathy, corneal and pressure and venous stasis ulcers, epidermolysis bullosa, acute myocardial infarction and one phase 1/2 in atherosclerotic cardiovascular disease - and not one of them lists arthritis of any kind, while a PubMed query pairing the molecule with arthritis returns ten records of which none is a treatment study. TB-500 is the seven-residue fragment LKKTETQ and is not the 43-residue thymosin beta-4 that almost all of that registered work used, so even the non-arthritis human record largely belongs to the parent protein. What the human rheumatoid arthritis literature does contain points the other way: thymosin beta-4 is raised in the synovial fluid and serum of patients and tracks disease activity, and a 2017 review in Biomedical Reports proposes testing whether antibodies AGAINST it relieve the disease.

Checked, and nothing recorded

12 compounds were checked against rheumatoid arthritis and produced no gradeable row.

• Balenine — checked, no row on file
• Bradykinin — checked, no row on file
• Chemerin — checked, no row on file
• Decapeptide-12 — checked, no row on file
• FOXO4-DRI — checked, no row on file
• GHK-Cu — checked, no row on file
• Omentin — checked, no row on file
• Palmitoyl Tetrapeptide-7 — checked, no row on file
• Psoriasin (S100A7) — checked, no row on file
• Selank — checked, no row on file
• Tetrapeptide-30 — checked, no row on file
• Thymulin — checked, no row on file

A compound is here because someone looked and the sources did not support a row — not because nobody looked. Under the rubric, uncertainty resolves to no row, and a row resting on a vendor page or a clinic blog is worse than no row at all.