Found in broccoli, cabbage, cauliflower, brussel sprouts, collard greens and kale.
Estimated DIM Exposure from Common Cruciferous Vegetables
| Vegetable |
Typical Serving |
Glucobrassicin / I3C Precursor |
Expected DIM Production |
Human Evidence |
Practical Interpretation |
| Brussels sprouts |
50–100 g |
High
~184 µmol glucobrassicin/100 g reported in a controlled human feeding study |
High relative to most common crucifers |
Strongest direct evidence. 50 g raw Brussels sprouts produced readily measurable urinary DIM. Individual 24-h urinary DIM ranged approximately 1.4–44 pmol/mg creatinine. |
Probably one of the most effective ordinary foods for producing I3C-derived DIM. Raw or lightly cooked sprouts that retain myrosinase are preferable. |
| Cabbage |
50–100 g |
Moderate
~46 µmol glucobrassicin/100 g in the controlled feeding study |
Low–moderate |
Direct human evidence. 50 g raw cabbage generated measurable DIM, but approximately several-fold less than the high-glucobrassicin Brussels sprouts used in the same trial. |
Effective dietary source, although cultivar strongly affects glucobrassicin concentration. |
| Broccoli |
75–100 g |
Moderate–high but extremely variable by cultivar
published glucobrassicin concentrations span a very wide range |
Moderate–high potential |
Glucobrassicin and I3C formation are well established, but precise DIM production after a normal broccoli serving has not been quantified as well as Brussels sprouts/cabbage. |
Potentially a very useful DIM-producing food. Broccoli is better known for glucoraphanin/sulforaphane, but some cultivars are also rich in glucobrassicin. |
| Cauliflower |
100 g |
Low–moderate
glucobrassicin is present but generally lower than in high-glucobrassicin Brussels sprouts |
Low–moderate potential |
Direct quantitative human DIM data are limited. |
Likely contributes meaningful I3C/DIM exposure but is not generally considered the richest dietary source. |
| Collard greens |
75–100 g |
Variable indole-glucosinolate content |
Low–moderate potential |
No good controlled human study directly quantifying DIM generated from a standard collard-green serving. |
Likely contributes to DIM exposure, but current evidence does not support assigning a precise mg DIM equivalent. |
| Kale |
75–100 g |
Moderate indole-glucosinolate content
~1 µmol/g dry weight glucobrassicin reported in some kale cultivars |
Low–moderate potential |
No strong human study directly quantifying DIM after a typical kale serving. |
Useful dietary source, although glucobrassicin content is highly cultivar-dependent and may be lower than Brussels sprouts. |
Important: vegetables do not normally contain large amounts of preformed DIM. They contain glucobrassicin, which is converted by myrosinase to indole-3-carbinol (I3C); I3C then undergoes acid condensation in the stomach to DIM and other oligomers. Therefore, glucobrassicin content is a better measure of dietary DIM-producing potential than the amount of DIM present in the vegetable itself.
Cruciferous Vegetable Exposure Compared with Research I3C Doses
| Source |
Typical Amount |
Approximate Glucobrassicin |
Maximum Theoretical I3C Equivalent |
Relative to 200 mg I3C |
Relative to 400 mg I3C |
Interpretation |
| Brussels sprouts |
50 g raw |
~92 µmol |
~13.5 mg |
~7% |
~3% |
Direct human feeding data; produces readily measurable urinary DIM. |
| Brussels sprouts |
100 g raw |
~184 µmol |
~27 mg |
~14% |
~7% |
A relatively high dietary exposure, but still substantially below purified-I3C clinical doses. |
| Cabbage |
50 g raw |
~23 µmol |
~3.4 mg |
~2% |
<1% |
Direct human data show measurable DIM formation despite the relatively low precursor dose. |
| Cabbage |
100 g raw |
~46 µmol |
~6.8 mg |
~3% |
~2% |
Useful dietary exposure but far below pharmacological I3C dosing. |
| Broccoli |
~100 g |
Highly cultivar-dependent |
Often roughly several to tens of mg |
Usually <10% |
Usually <5% |
Some high-glucobrassicin cultivars may approach Brussels sprouts; ordinary broccoli varies greatly. |
| Broccoli microgreens |
~25–50 g fresh |
Highly variable with cultivar and harvest age |
Usually low-to-tens of mg range |
Typically <10% |
Typically <5% |
Potentially concentrated relative to mature vegetables, but published values are too variable to assign a universal I3C equivalent. |
| Kale / collards / cauliflower |
~75–100 g |
Low to moderate and variable |
Usually several mg to low-tens of mg |
Usually <10% |
Usually <5% |
Meaningful dietary sources but generally not comparable with pharmacological I3C dosing. |
| Human vegetable dose-ranging study |
Cabbage + Brussels sprouts |
25–500 µmol |
~3.7–73.6 mg |
~2–37% |
~1–18% |
Urinary DIM increased with glucobrassicin dose, demonstrating biological uptake across this dietary range. |
| Purified I3C clinical trial |
200 mg/day |
Not applicable |
200 mg administered I3C |
100% |
50% |
CIN II–III trial; clinical lesion regression was reported in a small randomized study. |
| Purified I3C clinical trial |
300 mg/day |
Not applicable |
300 mg administered I3C |
150% |
75% |
Approximately the minimum dose associated with a significant estrogen-metabolism biomarker response in one dose-ranging study. |
| Purified I3C clinical trial |
400 mg/day |
Not applicable |
400 mg administered I3C |
200% |
100% |
Common human research dose producing substantial pharmacodynamic effects. |
| Purified I3C phase I |
800 mg/day |
Not applicable |
800 mg administered I3C |
400% |
200% |
Produced marked CYP1A2 induction; pharmacological rather than dietary exposure. |
Important: The theoretical I3C equivalents above are stoichiometric upper-bound comparisons, not measurements of absorbed I3C. Vegetable glucobrassicin must first be hydrolyzed by myrosinase, and I3C subsequently undergoes rapid gastric condensation into DIM and multiple other products. Therefore, 27 mg theoretical I3C from Brussels sprouts should not be interpreted as pharmacokinetically identical to swallowing 27 mg purified I3C.
Best approximate ranking for DIM production: Brussels sprouts > high-glucobrassicin broccoli > cabbage ≈ kale > cauliflower ≈ collard greens, but cultivar and preparation can change this order substantially.
Indole-3-carbinol — Indole-3-carbinol (I3C; indole-3-methanol; 3-hydroxymethylindole) is a naturally occurring indole phytochemical generated from the glucosinolate glucobrassicin when cruciferous vegetables are disrupted and plant myrosinase hydrolyzes the glucosinolate. It is classified as a dietary phytochemical and investigational chemopreventive agent rather than an approved anticancer drug. Major food sources include broccoli, Brussels sprouts, cabbage, cauliflower, kale, collards, and related Brassica vegetables. I3C is chemically unstable in gastric acid and rapidly forms multiple condensation products, particularly 3,3'-diindolylmethane (DIM); consequently, many systemic biological effects after oral I3C administration may actually be mediated by DIM and other acid-derived products rather than circulating parent I3C.
Primary mechanisms (ranked):
- Inhibition of proliferative and survival signaling, particularly PI3K/AKT and NF-κB, with downstream suppression of anti-apoptotic proteins and promotion of apoptosis.
- G1 cell-cycle arrest through modulation of cyclins, cyclin-dependent kinases, p21/p27, and retinoblastoma signaling.
- Activation/modulation of the aryl hydrocarbon receptor (AhR), including induction of CYP1-family enzymes and context-dependent pro-apoptotic transcriptional responses.
- Mitochondrial apoptosis involving ↑ Bax/Bcl-2 ratio, mitochondrial membrane disruption, cytochrome-c release, and caspase activation.
- Suppression of Wnt/β-catenin signaling, reducing proliferation, epithelial-mesenchymal transition, migration, and invasion in susceptible tumor models.
- Modulation of estrogen signaling and metabolism, including AhR-dependent ERα degradation and increased estrogen 2-hydroxylation.
- PTEN upregulation with suppression of tumor growth and enhancement of antitumor immune responses in recent colorectal-cancer models; this remains preclinical.
- Secondary/context-dependent oxidative-stress signaling, including ROS-dependent ATF3/NOXA-mediated apoptosis in some breast-cancer models; ROS modulation is not uniformly pro-oxidant across tissues or experimental systems.
Bioavailability / PK relevance: Oral I3C has unusual pharmacokinetics because acidic gastric conditions convert it rapidly into oligomeric products. In human pharmacokinetic studies, parent I3C was not detectable in plasma; DIM was the principal measurable circulating I3C-derived compound. After single oral doses of 400–1000 mg I3C, DIM exposure increased markedly, with an approximate Tmax of 2 hours, but exposure plateaued above about 1000 mg. Thus, oral I3C should be viewed pharmacologically as a precursor mixture that generates DIM and additional condensation products in the gastrointestinal tract rather than as a conventional systemically available parent compound.
In-vitro vs systemic exposure relevance: Many direct anticancer experiments expose cultured cells to I3C concentrations in the tens to hundreds of micromolar range, frequently around 100–300 µM. These concentrations substantially exceed plausible circulating parent-I3C exposure because parent I3C is generally undetectable after oral administration. Therefore, direct high-concentration I3C cytotoxicity in vitro has limited systemic PK relevance. Effects mediated by locally generated gastric products such as DIM, or by enzyme/receptor modulation occurring during gastrointestinal and hepatic exposure, are more biologically plausible after oral supplementation.
Clinical evidence status: Small human studies and early randomized trials are available, but there is no established clinical evidence that I3C treats invasive cancer. A small placebo-controlled randomized trial in cervical intraepithelial neoplasia II–III reported greater lesion regression with 200 or 400 mg/day I3C than placebo over 12 weeks. Phase-I studies in women found 400–800 mg/day generally tolerable and demonstrated substantial induction of CYP1A2 and altered estrogen metabolism. Evidence for established cancer therapy remains preclinical; recent reports of PTEN induction and enhancement of anti-PD-1 therapy are animal-model findings. I3C is not an approved anticancer therapy and should presently be categorized primarily as an investigational chemopreventive/pharmacologic dietary compound.
Indole-3-carbinol Cancer-Relevant Mechanisms
| Rank |
Pathway / Axis |
Cancer Cells |
Normal Cells |
Primary Effect |
Notes / Interpretation |
| 1 |
PI3K / AKT / NF-κB survival signaling |
↓ AKT
↓ NF-κB
↓ IKK
↓ anti-apoptotic signaling |
Less pronounced or ↔ in some non-tumorigenic epithelial models |
↓ survival
↑ apoptosis
↑ chemosensitivity |
One of the most repeatedly demonstrated I3C anticancer axes. Tumor-cell selectivity has been reported in paired malignant and nonmalignant breast epithelial models, but it is model-dependent. |
| 2 |
Cell-cycle control |
↓ cyclin D1
↓ cyclin E
↓ CDK2/CDK4/CDK6
↑ p21/p27
↓ Rb phosphorylation |
Usually less growth inhibition (model-dependent) |
G1 arrest
↓ proliferation |
G1 arrest is a consistent phenotype across several breast, prostate, leukemia, and other cancer models. |
| 3 |
Mitochondrial intrinsic apoptosis |
↑ Bax/Bcl-2 ratio
↓ mitochondrial membrane potential
↑ cytochrome-c release
↑ caspases |
Bax translocation may occur without full mitochondrial collapse or apoptosis in some normal epithelial cells |
↑ apoptotic cell death |
Provides evidence for partial tumor selectivity: mitochondrial depolarization and downstream apoptosis can be substantially stronger in malignant cells. |
| 4 |
AhR signaling |
↑ AhR activity
↑ CYP1A1
context-dependent ↑ apoptosis |
↑ xenobiotic-response enzymes |
Transcriptional reprogramming
chemoprevention or apoptosis |
AhR activation contributes directly to cytotoxicity in some colorectal and breast-cancer models. AhR biology is strongly tissue- and ligand-dependent and can also support tumorigenesis in other contexts. |
| 5 |
Wnt / β-catenin signaling |
↓ β-catenin
↓ c-Myc
↓ cyclin D1
↓ EMT |
Insufficient comparative evidence |
↓ proliferation
↓ migration
↓ invasion |
Supported by recent esophageal squamous-cell carcinoma experiments in vitro and xenografts and by earlier prostate-cancer studies. |
| 6 |
Estrogen receptor and estrogen metabolism |
↓ ERα (context-dependent)
↓ estrogen-dependent proliferation |
↑ CYP1A2/CYP1-family activity
↑ estrogen 2-hydroxylation |
Altered estrogen signaling and metabolism |
I3C can trigger AhR-dependent ERα degradation in ER-positive breast-cancer cells. Human supplementation substantially changes estrogen-metabolite ratios, but whether this reduces cancer incidence has not been established. |
| 7 |
PTEN / PI3K-AKT and antitumor immunity |
↑ PTEN
↓ proliferation/migration
↑ CD8+ T-cell infiltration |
Insufficient comparative evidence |
↓ tumor progression
↑ anti-PD-1 response |
Recent colorectal-cancer study demonstrated PTEN induction and potentiation of PD-1 antibody therapy in immunocompetent mice. This remains preclinical. |
| 8 |
Oxidative stress and ROS signaling |
↑ ROS (context-dependent)
↑ ATF3
↑ NOXA |
ROS may ↓ in other injury models |
Secondary promotion of apoptosis |
Not a universal I3C mechanism. Pro-oxidant signaling contributes to apoptosis in selected cancer models, whereas antioxidant and NRF2-related effects occur in other biological contexts. |
| 9 |
Migration and invasion |
↓ EMT
↓ migration
↓ invasion |
Insufficient comparative evidence |
↓ metastatic phenotype |
Likely downstream of several mechanisms including β-catenin, AKT/NF-κB, mitochondrial stress, and altered transcription rather than a single independent target. |
| 10 |
Chemosensitization and immunotherapy sensitization |
↑ sensitivity (agent- and model-dependent) |
Variable |
Potential adjunctive activity |
Preclinical enhancement has been reported with several cytotoxic drugs and, more recently, anti-PD-1 therapy. These combinations are not clinically validated. |
| 11 |
Clinical Translation Constraint |
High-concentration effects often difficult to reproduce systemically |
↑ CYP1A2 and other xenobiotic metabolism may alter drug exposure |
Limits direct translation of cell-culture results |
Parent I3C is acid-labile and essentially undetectable in human plasma after oral administration. DIM and other condensation products dominate systemic exposure. Many in-vitro I3C experiments use approximately 100–300 µM, making direct parent-compound cytotoxicity pharmacokinetically questionable. CYP1A2 induction creates a clinically relevant drug-interaction concern. |
|