Biotin Supplementation Ameliorates Murine Colitis by Preventing NF-Kb Activation Jonathan Skupsky,1,2 Subrata Sabui,3,4 Michael Hwang,3,5 Manando Nakasaki,6 Michael D

ORIGINAL RESEARCH Biotin Supplementation Ameliorates Murine Colitis by Preventing NF-kB Activation Jonathan Skupsky,1,2 Subrata Sabui,3,4 Michael Hwang,3,5 Manando Nakasaki,6 Michael D. Cahalan,4 and Hamid M. Said3,4,5

1Department of Medicine, Gastroenterology, 3Department of Medical Research, Veterans Affairs Long Beach, Long Beach, California; 2Department of Medicine, Division of Gastroenterology, 4Department of Physiology and Biophysics, 5Department of Medicine, 6Department of Pathology, University of California Irvine, Irvine, California

Active Colitis Activated Improved Colitis with NF-κB Inflammatory Biotin Therapy Cytokines

Lumen Intestinal Permeability Immune Infiltration

TNF-α

Mucosa IL-1β Biotin IL-6

fl SUMMARY length,aswellasexpressionofinammatory cytokines (interleukin 6, tumor necrosis factor-a, interleukin 1b), in- Biotin deficiency causes an inflammatory bowel disease–like testinal permeability, tight junctions (zonula occludens-1 and k state, but biotin supplementation has never been tested as a claudin-2), and the transcription factor nuclear factor- B k therapeutic in established models for inflammatory bowel (NF- B). disease. We found that biotin therapy leads to delayed onset RESULTS: Biotin therapy led to delayed onset and severity of and severity of dextran sodium sulfate colitis mediated by colitis as well as accelerated healing. There was improvement k decreased activation of nuclear factor- B. in the Disease Activity Index, fecal calprotectin levels, colon length, and histology. In addition, biotin-treated mice had reduced expression of inflammatory cytokines, reduced intes- BACKGROUND & AIMS: Biotin is a water-soluble vitamin that tinal permeability, and reduced activation of NF-kB. is indispensable for human health. Biotin deficiency can CONCLUSIONS: Oral supplementation with biotin provides cause failure-to-thrive, immunodeficiency, alopecia, derma- benefit for maintenance and induction of remission in the titis, and conjunctivitis. We previously reported that biotin dextran sodium sulfate preclinical model for IBD. Biotin does deficiency also can lead to severe colitis in mice, which is this by reducing the activation of NF-kB, which prevents the completely reversed with supplementation. Our aim in this production of inflammatory cytokines and helps maintain the study was to determine if high-dose biotin supplementation integrity of the intestinal barrier. Clinically, the NF-kB pathway can provide a therapeutic benefit in a preclinical model for is important in the development of IBD and this finding sug- inflammatory bowel disease (IBD) and to identify the mo- gests that biotin may have therapeutic potential for patients lecular mechanism by which this occurs. with IBD. (Cell Mol Gastroenterol Hepatol 2020;9:557–567; METHODS: Mice were challenged with dextran sodium sul- https://doi.org/10.1016/j.jcmgh.2019.11.011) fate to induce colitis and were treated with 1 mmol/L biotin to induce or maintain remission. Clinical response was Keywords: Biotin; Inflammatory Bowel Disease; Colitis; monitored by the Disease Activity Index and fecal calpro- Therapeutics. tectin levels. The colon tissue was investigated for histology, 558 Skupsky et al Cellular and Molecular Gastroenterology and Hepatology Vol. 9, No. 4

depending on the side-effect profile, patients will need to be iotin is a water-soluble vitamin and an essential monitored for adverse events.26 We propose that biotin micronutrient that must be obtained from exogenous B would be an ideal medication for IBD therapy because it can sources such as dairy, liver, egg yolk, and vegetables, or be delivered orally and has minimal risk for complications. from commensal bacteria.1 Its classic role is as a covalently Notably, biotin deficiency often is overlooked in the setting bound coenzyme for cytoplasmic carboxylases used in fatty of IBD and there have been several reports of biotin defi- acid homeostasis, gluconeogenesis, and other metabolic – ciency in patients with IBD.27 31 The data from mice treated pathways,2 but emerging data also have identified a role in with biotin to prevent or treat colitis suggest that biotin cellular stress response,3 gene regulation,4,5 and immune – therapy may provide benefit to patients with IBD. responses.6 10 Clinically, biotin deficiency, whether induced by dietary means or by mutation of the intestinal biotin uptake system (ie, sodium-dependent multivitamin trans- Results porter [SMVT], Slc5a6), is associated with failure to thrive, The Dietary Model for Biotin Deficiency Induces microcephaly, osteoporosis, immunodeficiency, alopecia, an IBD-Like State, Which Is Prevented With 11,12 dermatitis, and conjunctivitis. In our previous studies to Biotin Supplementation determine the relative contribution of SMVT toward total In the first series of experiments, we examined dietary carrier-mediated biotin absorption, we made the serendip- biotin deficiency to further characterize the phenotype that fi itous nding that SMVT knockout mice also develop colonic develops. Mice received a biotin-deficient diet and showed crypt abscesses, neutrophil infiltration, submucosal edema, fi 13 the rst signs of disease after 7 weeks when they began to and dysplastic changes. In the current report, we extend develop alopecia and weight loss. As time progressed, they fi fi our ndings on the effect of biotin de ciency and show that had decreased activity levels, hunched posture, poor skin fi dietary-induced biotin de ciency leads to an increase in the turgor, and their stools became soft and bloody. This was fl in ammatory marker fecal calprotectin, which is used prevented with 1 mmol/L biotin supplementation added to routinely to monitor the progression of patients with in- their drinking water (Figure 1A and B). We further tested fl 14 ammatory bowel disease (IBD). Importantly, we show stool samples at week 14 for fecal calprotectin, a noninva- that these mice fully recover from their IBD-like state with sive biomarker for inflammation in the gastrointestinal (GI) biotin supplementation. tract, and found that it was increased in mice with biotin Our goal in this study was to determine if biotin sup- deficiency (Figure 1C). Inflammation was most pronounced plementation also could be used as treatment in a pre- in the cecum of mice with biotin deficiency and histology clinical model for IBD. For this, we examined the most showed loss of goblet cells, altered crypt architecture, widely used experimental model for IBD, the dextran so- fi 15,16 in ltration of lymphocytes and neutrophils, with expansion dium sulfate (DSS) colitis model. Mice were challenged of the lamina propria. Mice treated with biotin supplemen- with DSS in drinking water to induce colitis; after 7 days tation had normal histology without colitis (Figure 1D). they had increased weight loss, Disease Activity Index Although no mouse model entirely recapitulates patients fl (DAI) scores, calprotectin, proin ammatory cytokines, in- with IBD, this model reproduces many of the findings testinal permeability, and histologic scores, as expected. In including weight loss, bloody diarrhea, increased fecal cal- addition, they also had decreased levels of the biotin protectin, altered crypt architecture, and infiltration of fi transporter SMVT. The clinical relevance of this nding was neutrophils and lymphocytes to the mucosa and submucosa. confirmed when we identified that there also was a decrease of SMVT in biopsy specimens from patients with active ulcerative colitis (UC). Next, we used the DSS model The Biotin Transport Pathway Is a Clinically to examine the ability of biotin supplementation to Relevant Target for IBD Treatment ameliorate signs of colitis and found that all parameters Because biotin therapy was able to reverse the IBD-like measuring the development of colitis trended toward features described earlier, the next series of experiments baseline with biotin therapy. Finally, we identified the were designed to validate if the biotin transport pathway mechanism by which biotin exerts its therapeutic effect by would be a relevant drug target in IBD. To address a role for showing decreased levels of active phosphorylated nuclear biotin in the treatment of colitis, we induced severe disease factor-kB(NF-kB) in biotin-treated mice. NF-kBisapleio- in mice using 3% DSS in drinking water. The mice in this tropic transcription factor that drives expression of pro- survival genes in intestinal epithelial cells, and it also Abbreviations used in this paper: DAI, Disease Activity Index; DSS, coordinates expression of proinflammatory genes from dextran sodium sulfate; ELISA, enzyme-linked immunosorbent assay; cells of the innate and adaptive immune system.17,18 FITC, fluorescein isothiocyanate; GI, gastrointestinal; IBD, inflamma- tory bowel disease; IL, interleukin; NF-kB, nuclear factor-kB; PCR, Increased inflammation promotes epithelial permeability, polymerase chain reaction; SMVT, sodium-dependent multivitamin which exacerbates colitis. Indeed, NF-kB is a pivotal regu- transporter; TNF, tumor necrosis factor; UC, ulcerative colitis. 19,20 lator in the development of IBD. Most current article In recent years, several new medications have been © 2020 The Authors. Published by Elsevier Inc. on behalf of the AGA approved, and many more are under development, for the Institute. This is an open access article under the CC BY-NC-ND ’ 21–25 license (http://creativecommons.org/licenses/by-nc-nd/4.0/). treatment of UC and Crohn s disease. For each medi- 2352-345X cation, there is a risk-to-benefit evaluation to consider and, https://doi.org/10.1016/j.jcmgh.2019.11.011 2020 Biotin Supplementation Ameliorates Colitis 559

A Control BD Diet + B Diet BD Diet 1mM Bion Percent Starng Weight Percent Starting Weight Percent Starting

0 1 2 3 4 5 6 7 8 9 4 10 11 12 13 1 Weeks on Diet BD Diet + CDControl Diet BD Diet 1mM Bion ► ► →

(ng/g stool) →

Fecal Calprotecn →

Figure 1. Chronic biotin deficiency induces an IBD-like state that completely resolves with biotin supplementation. C57BL/6J mice received a control diet or a biotin-deficient diet with and without 1 mmol/L biotin supplementation for 14 weeks. (A) Representative images comparing a control mouse, a mouse on the deficient diet ,and a mouse that received the deficient diet and treatment with 1 mmol/L biotin. (B) Body weight was recorded weekly for mice in each group, significance was calculated relative to the BD diet group. (C) A stool pellet was collected from mice in each group the day before death and used to quantify the fecal calprotectin level. (D) Representative histology from the cecum of a mouse in each group. Mice with biotin deficiency have erosions (arrowheads), loss of goblet cells, and altered crypt architecture with increased infiltration of lymphocytes and neutrophils into the lamina propria (arrows), and expansion of the submucosa, while mice that were sup- plemented with biotin did not have cecal inflammation. *P < .05, **P < .01, and ***P < .001. BD, biotin deficiency. model typically develop a robust response within 7 days colon and this raises the possibility that it could be a target that can be measured by standard parameters including for therapy. DAI, colon length, and histologic scoring (Figure 2A–E). We wanted to study the biotin transport pathway in this acute model for colitis and found that the localized biotin trans- Biotin Therapy Helps Maintain Remission in DSS port pathway in the colon was down-regulated, as judged Colitis from protein isolated from the distal colon to quantify To determine if biotin can be used as a therapeutic for expression of SMVT (Figure 2F). These results suggested the maintenance of remission to colitis, we used a 1.5% DSS that physiologic biotin uptake may be limited in the setting model to induce mild/moderate colitis. Starting with healthy of colitis. mice, 1 group received 1 mmol/L biotin in their drinking To determine if there is clinical relevance of these find- water for 1 week before DSS challenge. On day 0, the ings, we quantified SMVT expression by real-time poly- experimental and treatment groups had DSS added to merase chain reaction (PCR) using tissue from patients with drinking water and the treatment group also was continued colitis and controls. We found that SMVT was decreased on biotin (Figure 3A). An additional control group did not significantly in patients with moderate/severe UC compared receive biotin or DSS. We found that disease in the biotin with controls (Figure 2G). To extend this finding, we treatment group was attenuated in comparison with the reviewed a larger, publicly available gene expression data- mice without supplementation. In addition, the onset of base32 (data accessible at NCBI GEO database, accession disease was delayed, and the peak of disease was decreased GDS3268) with 202 biopsy specimens taken from patients with biotin therapy, as measured by DAI (Figure 3B). The with controlled UC, active UC, and healthy controls.33 We colons from DSS challenged mice contained blood and were searched for expression levels of the SMVT gene, Slc5a6,in shortened significantly compared with controls and those the sigmoid colon and the data have been normalized to that received DSS with biotin therapy. There was no dif- samples from patients without disease. Interestingly, SMVT ference between mice that received biotin therapy and is reduced significantly in patients with inflamed UC, but controls lacking DSS exposure (Figure 3C and D). Fecal there is no difference between healthy controls and patients calprotectin also was quantified by enzyme-linked immu- with uninflamed UC (Figure 2H). Overall, these findings nosorbent assay (ELISA) and found to be increased in the show that in mice with DSS colitis and in human beings with DSS group compared with controls and mice that received active UC, the biotin transport pathway is altered in the DSS with biotin therapy. Again, there was no significant 560 Skupsky et al Cellular and Molecular Gastroenterology and Hepatology Vol. 9, No. 4

AB C *** *** ***

*** *** Disease Acvity

* (cm) Length Colon Figure 2. SMVT expres- Colon Length (cm) Index (DAI, max. (DAI, 4) Index sion was reduced in hu- man beings and mice with colitis. (A) Experi- D E mental protocol. (B) DAI No DSS 3% DSS was calculated daily. (C) *** 8 On day 7, the mice were killed and colon length 6 was determined. (D) Representative histology

4 from the distal colon of mice treated with DSS and (max. 9) controls. DSS-treated 2 → → Histology Score mice developed mucosal erosion (arrowheads), with 0 → No DSS 3% DSS loss of goblet cells, altered crypt architecture, and infiltration of neutrophils F G and lymphocytes (arrows). (E) Total histologic score. Control 3% DSS (F) Protein was isolated SMVT from a segment of the distal colon and SMVT β-acn expression was quantified by Western blot and b

Relative mRNA normalized to -actin and Relave mRNA Relave Protein the healthy control. (G) Slc5a6 levels were quanti- fied by quantitative PCR from patients with moder- H ate/severe left-sided colitis and normalized to b-actin and healthy controls. (H) Slc5a6 levels from the sigmoid colon in patients with UC was normalized to healthy controls using data available from GEO

SMVT Expression SMVT accession GDS3268. *P < SMVT Expression .05, **P < .01, and ***P < .001. max, maximum; mRNA, messenger RNA.

difference between the group that received biotin therapy Biotin Therapy Enhances Induction of Remission and the no-DSS control (Figure 3E). Finally, histology to DSS Colitis showed mucosal erosion, distorted crypts, loss of goblet The next set of experiments were designed to determine if cells, and infiltration of neutrophils and lymphocytes in the there is a role for biotin therapy in the induction of remission mucosal and submucosal layers of the colon wall of mice to colitis. Mice received 1.5% DSS in their drinking water for 7 that received DSS. In contrast, mice that received biotin days to induce moderate colitis. On day 0, the DSS was treatment maintained much of the mucosal lining with well- removed and half the mice received water, while the other half preserved crypts and goblet cells, and they had only mild received therapy with 1 mmol/L biotin. Again, there was a infiltration into the lamina propria (Figure 3F). Histologic control group that did not receive biotin or DSS. The groups scoring confirmed significant improvement in colitis with were monitored until day 4 when the experiment was ended biotin therapy (Figure 3G). Overall, these data show that and all mice were killed (Figure 4A). Mice that received biotin biotin therapy helps to attenuate the development of colitis. therapy recovered more quickly with complete resolution of 2020 Biotin Supplementation Ameliorates Colitis 561

A B

Figure 3. Biotin therapy helps maintain remission and ameliorate the development of colitis. Disease Acvity

(A) Experimental protocol. max. (DAI, 4) Index (B) DAI was calculated daily and significance is shown for biotin treatment. (C) On day 7, the mice C D E were killed and a repre- 8000 ** sentative image of a colon ** * from each group is shown. 6000 (D) Average colon length for each group. (E) A stool 4000 pellet was collected from (ng/g stool) mice in each group before 2000 Fecal Calprotecn killing and used to quantify Colon Length (cm) the fecal calprotectin level. 0 (F) Representative histol- No DSS 1.5% DSS 1.5% DSS + No DSS 1.5% DSS 1.5% DSS + 1mM Biotin ogy from the distal colon 1mM Biotin of a mouse in each group. F 1.5% DSS + G DSS-treated mice had No DSS 1.5% DSS 1mM bion mucosal erosion (arrow- heads) and substantial infiltration of neutrophils and lymphocytes in the lamina propria with loss of → crypts and goblet cells → (arrows). (G) Total histo- → < < logic score. *P .05, **P Histology Score (max. 9) .01, and ***P < .001. max, maximum.

disease, as measured by DAI (Figure 4B). The colons from mice ameliorate colitis in the models for both maintenance and that had received only water after DSS appeared shortened induction of remission. We previously have seen that in- and atrophic compared with mice that had received biotin flammatory cytokines are up-regulated in the setting of therapy and the no-DSS control. Nearly all the mice in all biotin deficiency34 and that those cytokines can be groups had solid stools by that point, although the average normalized with biotin supplementation.35 To determine if colon length still was shortened in the group that had received there were similarities between the biotin deficiency and only water after DSS. There was no difference between mice the DSS colitis models, we examined 3 proinflammatory that received biotin induction therapy and controls (Figure 4C cytokines that have a pivotal role in the pathogenesis of IBD: and D). The fecal calprotectin level was increased in the group interleukin (IL)6, tumor necrosis factor (TNF)a, and that received only water after DSS, compared with mice that IL1b.36–38 Samples from the distal colons of mice in the received biotin therapy and controls. Again, there was no maintenance of remission experiment were assayed by significant difference between the group that received biotin quantitative real-time-PCR. IL6 levels were increased therapy and the control group (Figure 4E). Histology in mice significantly in mice that received DSS, although they were from both groups showed marked improvement from the peak near baseline with biotin supplementation. TNFa and IL1b of disease, but the group that received only water still had mild levels also were increased with DSS colitis, but cytokine crypt loss and neutrophil and lymphocyte infiltration levels did not completely return to baseline levels with (Figure 4F and G). Overall, these data indicate that biotin biotin therapy (Figure 5A). There are similar trends for mice therapy is able to accelerate healing during induction of during the induction of remission (Figure 5B). These data remission. complement earlier studies showing that biotin levels can affect innate and adaptive immune responses.6,8–10,13,35,39 Biotin Therapy Leads to a Reduction in To address whether or not biotin therapy alters mucosal Inflammatory Cytokines and Intestinal permeability we performed a fluorescein isothiocyanate Permeability (FITC)-dextran permeability assay after the induction of We next examined localized cytokine production to remission protocol. We noted increased levels of FITC- investigate the mechanism by which biotin was able to dextran in the plasma of mice that had received only water 562 Skupsky et al Cellular and Molecular Gastroenterology and Hepatology Vol. 9, No. 4

A B 2.5 Water 1mM Biotin 2.0 * ** No DSS

1.5 ** 1.0 *

Disease Acvity 0.5 Index (DAI, max. (DAI, 4) Index Figure 4. Biotin therapy 0.0 0 1 2 3 4 accelerates induction of Day remission to DSS colitis. C D (A) Experimental protocol. N/S E (B) DAI was calculated 8 * * daily and significance is shown for biotin treatment. 6 (C) On day 4, the mice were killed and a repre- 4 sentative image of a colon

(ng/g stool) from each group is shown. 2 Colon Length (cm)

Fecal Calprotecn (D) Average colon length. (E) A stool pellet was No Water 1mM 0 collected from mice in Biotin No DSS Water 1mM Biotin DSS each group before killing and used to quantify the F G fecal calprotectin level. (F) No DSS Water 1mM bion Representative histology from the distal colon of a → mouse in each group. Mice that received only → water continued to have moderate infiltration along → with loss of crypts and goblet cells (arrows). (G) Total histologic score. *P < .05 and **P < .01. max, maximum.

after DSS exposure, whereas mice that had received biotin unstimulated cells. Activation signals allow for nuclear therapy showed similar values to the controls that received translocation of the dimers and subsequent transactivation no DSS (Figure 6A). These data show that biotin therapy of target genes.20,43 The p65 protein is the most abundant helped reduce intestinal permeability and the results were and best studied in the NF-kB family, and phosphorylation is confirmed by examining the expression of tight junction the starting point in its activation.43,44 We found that total proteins. In the intestine, up-regulation of zonula occludens-1 NF-kB p65 levels were unchanged with biotin therapy is important to maintain the mucosal integrity,40 while up- (Figure 7A), while the phosphorylated active form (NF-kB regulation of claudin-2 is associated with susceptibility to p65 [phospho S529]) was increased in mice that did not colitis and increased intestinal permeability.41,42 Zonula receive biotin supplementation (Figure 7B). These results occludens-1 levels in the distal colon were decreased in mice show that biotin therapy attenuates activation of NF-kB p65 that had received only water after DSS, while they were close and the inflammatory cascade that follows. to baseline with biotin therapy (Figure 6B). Conversely, claudin-2 expression was increased in the distal colons of Discussion mice that had received only water after DSS while they were In this report, we showed that the colitis that develops again close to baseline with biotin therapy (Figure 6C). from biotin deficiency has many similarities to mouse models Overall, these data support a model in which biotin supple- for IBD including weight loss, bloody stool, increased fecal mentation helps to maintain mucosal integrity. calprotectin, and infiltration of lymphocytes and neutrophils to the intestinal wall. Importantly, disease and this IBD-like Biotin Therapy Reduces NF-kB Activation state can be prevented with biotin supplementation. Biotin Finally, we wanted to explore these changes at the mo- is one of the water-soluble vitamins with transporters in both lecular level to determine if NF-kB is affected by biotin the small and large intestines.45 The current understanding is therapy. NF-kB is considered one of the primary regulators that dietary biotin is absorbed primarily in the small intestine in the development of IBD.20 The 5 proteins in the NF-kB and biotin produced by commensal bacteria is absorbed in family dimerize and localize to the cytoplasm in the large intestine. Because the colon has constant exposure 2020 Biotin Supplementation Ameliorates Colitis 563

A IL-6

500 N/S

400 * *

300

200 Figure 5. Biotin therapy Relative mRNA Relative leads to reduced proin- 100 Relative mRNA flammatory cytokines in 0 the colon. (A) Quantitative No DSS 1.5% DSS 1.5% DSS real-time-PCR was per- + 1mM Biotin formed on tissue from the distal colon to quantify the cytokines IL6, TNFa, and B IL1b from mice in the maintenance of remission protocol, and (B) the in- duction of remission pro- tocol. Data have been normalized to glyceralde- Relative mRNA Relative hyde-3-phosphate dehy- Relative mRNA drogenase and healthy controls. *P < .05 and **P < .01. mRNA, messenger RNA.

to biotin, it absorbs a substantial amount of total biotin.46 result of a colitis flare, it does suggest that biotin supple- These observations led us to evaluate if the biotin transport mentation should be explored for IBD treatment. pathway is affected in an established mouse model for IBD. We then tested biotin supplementation in the DSS pre- We used DSS to induce severe colitis, and examined mice at clinical model for IBD to determine if it has a therapeutic ef- the peak of disease. We found that SMVT levels in the distal fect. There is no expected direct interaction between biotin colon, where inflammation is most pronounced in the DSS and DSS.2 In the maintenance of remission model, mice began model, were reduced significantly in mice with colitis when without disease and received biotin supplementation. They compared with healthy controls. The clinical relevance of this then were challenged with DSS to induce moderate colitis and finding was supported by comparing gene expression in pa- we monitored the changes between the treatment and control tients with UC and healthy controls. We found decreased groups for 7 days. The mice receiving biotin therapy still expression of the SMVT biotin transporter in the distal colon developed some signs of disease, but they were substantially of patients with active UC when compared with the distal less severe in comparison with mice that did not receive colon of healthy controls. The same finding of decreased biotin. In biotin-treated mice, DAI showed a delayed onset and SMVT expression during active disease can be found in the a reduced peak of disease, fecal calprotectin level was proximal and descending colon as well.32 Although this as- reduced, and histology showed only mild lymphocyte infil- sociation cannot determine if reduced SMVT is a cause or tration compared with more pronounced erosions and crypt

A N/S B C 2000 * *

1500

1000

500

0 No DSS Water 1mM Biotin

Figure 6. Biotin therapy helps maintain the integrity of the intestinal mucosa. (A) Mice from the induction of remission protocol received FITC-dextran by oral gavage 4 hours before killing. Plasma levels of FITC-dextran were measured as a marker of intestinal permeability. (B) Quantitative real-time-PCR was performed on tissue from the distal colon to quantify tight junction protein zonula occludens-1 (ZO-1) and (C) claudin-2. Data have been normalized to glyceraldehyde-3-phosphate dehydrogenase and healthy controls. *P < .05 and **P < .01. mRNA, messenger RNA. 564 Skupsky et al Cellular and Molecular Gastroenterology and Hepatology Vol. 9, No. 4

dual role in the GI tract by driving proinflammatory responses A in immune cells and by up-regulating prosurvival genes in 17 No DSS + 1mM epithelial cells. However, in the setting of IBD the primary DSS DSS Bion role is orchestrating mucosal inflammation.20 In the current NF-κB p65 study, we showed that biotin therapy is sufficient to attenuate k β-acn activation of NF- B p65 and propose this as the underlying mechanism of its therapeutic effect. If infiltrating cells have reduced activation of NF-kB, then they will produce decreased Relative protein amounts of inflammatory cytokines that subsequently will lead to reduced intestinal permeability and maintenance of mucosal integrity. The specific mechanism(s) linking biotin and NF-kB is unclear but could be mediated via the different cellular pathways that are affected by biotin availability. B Further studies are required to define this relationship. No DSS + 1mM IBD develops from a combination of genetic predisposi- DSS DSS Bion tion, environmental triggers, dysbiosis in the microbiome, NF-κB p65 (phospho S529) and complex interactions in the host immune system, making it challenging to model IBD accurately in preclinical β-acn models.50,51 As mentioned, biotin deficiency has been identified in patients with IBD and it may have profound

Relative protein implications in the GI tract that have not been fully explored. We are optimistic that the data presented here will serve as the foundation for future clinical studies to determine if biotin supplementation should be used as adjunct therapy in IBD. Biotin is available over the counter, is affordable, and it Figure 7. Biotin therapy leads to decreased levels of has minimal side effects, making it an ideal therapeutic if active phosphorylated NF-kB p65. Protein was isolated clinical trials can show similar efficacy to what we have seen k from a segment of the distal colon. (A) NF- B p65 and (B) in this preclinical model. phosphorylated NF-kB p65 (phospho S529) expression was quantified by Western blot and normalized to b-actin and the healthy control. Data shown are the means ± SE. Statistical Materials and Methods analyses were performed with a Student t test. *P < .05 and All authors had access to the study data and reviewed < ***P .001. and approved the final manuscript. loss in the untreated mice. Proinflammatory cytokines IL6, TNFa, and IL1b also were decreased in the treatment group. Mice We repeated a similar set of experiments using an induction of C57BL/6J mice were purchased from Jackson Labora- remission model. In this model, mice began biotin therapy at tories (Ellsworth, ME) and housed at University of California the peak of disease and we compared remission rates with Irvine in compliance will all University Laboratory Animal – untreated mice after the DSS was removed. After 4 days on Resources guidelines. Mice aged 8 10 weeks were used at biotin, the DAI score of treated mice had returned to baseline the start of all experiments and representative experiments fi and was improved significantly in comparison with untreated were con rmed with male and female mice to ensure there mice. Similar results were seen for colon length, calprotectin, were no sex differences. Experiments were performed in and histology. Again, proinflammatory cytokines were compliance with all federal and local guidelines and decreased. In addition, in this model, we evaluated gut approved by the Institutional Animal Care and Use Com- permeability and found that it was decreased with biotin mittee and the University of California Irvine. treatment and that there were increased levels of zonula fi occludens-1 and decreased levels of the leaky tight-junction Biotin-De cient and Control Diet fi protein claudin-2. These data support our previous finding The biotin-de cient diet (Envigo, Madison, WI) has no that the biotin transport pathway plays an important role in added biotin and the protein source is 300 g/kg spray-dried, the maintenance of mucosal integrity.34 egg-white solids. Egg white is high in avidin, which bind the 7,52 The mechanisms underlying the pathogenesis of biotin luminal biotin produced by the host microbiome. The deficiency and biotin’s therapeutic potential remains an active biotin-control diet (Envigo) has the same make-up and fi area of research and several advances have been made protein source as the de ciency diet, but is supplemented – recently.6,8,13,34,35,39,47 49 We previously reported on an SMVT with 0.004 g/kg biotin. Mice received the diet ad libitum and knockout mouse that provided evidence that the SMVT system pair-feeding was performed. is exclusively responsible for intestinal biotin uptake and deletion of SMVT is sufficient to induce colitis.2 SMVT levels Fecal Calprotectin ELISA also can decrease physiologically in response to cytokine Fecal calprotectin was quantified by ELISA using the activation of the NF-kBpathway.39 Interestingly, NF-kBhasa DuoSet Mouse S100A8/S100A9 Heterodimer kit (R&D 2020 Biotin Supplementation Ameliorates Colitis 565

Table 1.List of Primer Sequences Used for Quantitative Real-Time-PCR

Gene name Forward and reverse primer sequences, 50-30 Mouse claudin-2 TTAGCCCTGACCGAGAAAGA; AAAGGACCTCTCTGGTGCTG Mouse ZO-1 TTCAAAGTCTGCAGAGACAATAGC; TCACATTGCTTAGTCCAGTTCC Mouse IL6 GAGGATACCACTCCCAACAGACC; AAGTGCATCATCGTTGTTCATACA Mouse TNF-a CATCTTCTCAAAATTCGAGTGACAA; TCGGAGTAGACAAGGTACAACCC Mouse IL1b CTCTCCAGCCAAGCTTCCTTGTGC; GCTCTCATCAGGACAGCCCAGGT Mouse GAPDH CTACAGCAACAGGGTGGTGG; TATGGGGGTCTGGGATGG Mouse SMVT CGTAGGAACTTTGGTAGCCCTGG; CTTAGGTGTGATGGGTCTCTCC Human SMVT TGTCTACCTTCTCCATCATGGA; TAGAGCCCAATGGCAAGAGA Human b-actin CATCCTGCGTCTGGACCT; TAATGTCACGCACGATTTCC

GAPDH, glyceraldehyde-3-phosphate dehydrogenase; ZO-1, zonula occludens-1.

Systems, Minneapolis, MN) according to the manufacturer’s body weight. Mice were fasted overnight and gavaged with recommendations. Briefly, a stool pellet was collected, FITC-dextran on the last day of the experiment. After 4 hours, weighed, and immediately frozen. The night before the assay, mice were bled, plasma was separated, and measurements pellets were incubated at 4C with 1 mL fecal extraction were taken by fluorimeter at 488 nm (Synergy BioTek Plate buffer (0.1 mol/L Tris, 0.15 mol/L MaCl, 1.0 mol/L urea, Reader, Winooski, VT). The concentration was calculated 10 mmol/L CaCl2, 0.1 mol/L citric acid, 5 g/L bovine serum based on a stander curve generated by serial dilutions of the albumin, 0.25 mmol/L thimerosal, pH 8; Hycult Biotech, FITC-dextran stock in phosphate-buffered saline. Wayne, PA). The supernatant was collected after the pellet was broken up with vortexing and ceramic beads. The dilu- Quantitative Real-Time PCR Analysis tion with extraction buffer was calculated assuming a density Total messenger RNA was isolated from the mouse distal of 1 mg/mL and results were normalized to the stool weight. colon with Qiazol and the RNeasy Kit (Qiagen, Hilden Ger- many) using 1.2-mm silica beads (Fisherbrand, Hampton, Histologic Analysis NH). To remove DSS from purified messenger RNA we De-identified H&E-stained sections from DSS-treated performed further messenger RNA purification by lithium mice were scored by a pathologist according to published chloride as described.56 Complementary DNA was synthe- methods.53 Briefly, the score is the sum of neutrophil infil- sized using the Verso complementary DNA synthesis kit tration (0–2), lymphocyte infiltration (0–2), erosion (0–3), (Invitrogen, Carlsbad, CA). Expression levels of human and crypt loss (0–2), with a maximum score of 9. SMVT was determined using formalin-fixed paraffin- embedded biopsy samples collected from the sigmoid colon DSS Colitis of patients at University of California Irvine with moderate/ fi DSS salt (molecular weight, 36,000–50,000; MP Bio- severe UC. RNA was puri ed from the samples using the fi medicals, Irvine, CA) was added to drinking water for Quick-RNA Formalin-Fixed Paraf n-Embedded Kit (Zymo fi 1 week at 3% to induce severe colitis or 1.5% for mild/ Research, Irvine, CA). Samples were quanti ed by real-time fi moderate colitis. The solution was changed every 2–3 days. quantitative PCR using gene-speci cprimers(Table 1) with Mice were observed daily and the DAI was calculated ac- a CFX96 real-time iCycler (Bio-Rad, Hercules, CA). SYBR cording to published protocols.54 Briefly, the score given is green was used for mouse complementary DNA samples, the average of a score for weight loss (0, none; 1, 2%–5%; 2, and the iTaq Universal One-Step Kit (Bio-Rad) was used for 5%–10%; 3, 10%–15%; 4, >15%), stool consistency human RNA samples. (0, normal; 2, loose stool; 4, diarrhea), and blood in stool (0, negative; 2, guaiac positive, 4; gross bleeding). Western Blot For Western blot analysis, tissues/cells were homoge- nized in RIPA buffer (Sigma) containing complete protease Biotin Therapy inhibitor cocktail (Roche, Basel, Switzerland) using 1.2-mm Biotin (Sigma-Aldrich) was dissolved in drinking water to silica beads (Fisher, Hampton, NH) and the Fisherbrand a concentration of 1 mmol/L. The solution was changed Bead Mill 24 Homogenizer. Total protein homogenates were – every 2 3 days and mice received the treatment ad libitum. cleared by centrifugation at 12,000g for 20 minutes, and an equal amount (65 mg) of the total proteins was loaded on a FITC Dextran Permeability Assay 4%–12% mini gel (Invitrogen). The proteins then were Intestinal permeability was determined in vivo using transferred to a polyvinylidene difluoride membrane and methods previously described.55 Briefly, FITC-dextran (FD4; probed simultaneously with anti-mouse SMVT, NF-kB, Sigma-Aldrich) was dissolved in phosphate-buffered saline phospho NF-kB p65 (ser536) (raised in rabbit), and mono- to 100 mg/mL and administered to mice at 44 mg/100 g clonal b-actin antibody (raised in mouse). The blots then 566 Skupsky et al Cellular and Molecular Gastroenterology and Hepatology Vol. 9, No. 4 were incubated with anti-rabbit/anti-mouse IR 800 dye and 14. Brookes MJ, Whitehead S, Gaya DR, Hawthorne AB. anti-mouse IR 680 dye (LI-COR, Lincoln, NE) secondary Practical guidance on the use of faecal calprotectin. antibodies (1:25,000) for 1 hour at room temperature. Frontline Gastroenterol 2018;9:87–91. Relative expression was quantified by comparing the fluo- 15. Eichele DD, Kharbanda KK. Dextran sodium sulfate colitis rescence intensities in an Odyssey Infrared imaging system murine model: an indispensable tool for advancing our (LI-COR) using Odyssey application software (version 3.0) understanding of inflammatory bowel diseases patho- with respect to corresponding b-actin. genesis. World J Gastroenterol 2017;23:6016–6029. 16. Okayasu I, Hatakeyama S, Yamada M, Ohkusa T, Statistical Analysis Inagaki Y, Nakaya R. A novel method in the induction of The data are expressed as means ± SEM. Differences reliable experimental acute and chronic ulcerative colitis – were analyzed by a Student t test and was considered sig- in mice. Gastroenterology 1990;98:694 702. nificant when the P value was less than .05. Experiments 17. Kondylis V, Kumari S, Vlantis K, Pasparakis M. The were performed at least 3 times to confirm consistent re- interplay of IKK, NF-kappaB and RIPK1 signaling in the fl sults and data from the most recent experiment are shown. regulation of cell death, tissue homeostasis and in am- mation. Immunol Rev 2017;277:113–127. 18. Zhang Q, Lenardo MJ, Baltimore D. 30 years of NF- References kappaB: a blossoming of relevance to human pathobi- 1. Said HM. Water-soluble vitamins. World Rev Nutr Diet ology. Cell 2017;168:37–57. 2015;111:30–37. 19. Li Q, Verma IM. NF-kappaB regulation in the immune 2. Said HM. Biotin: biochemical, physiological and clinical system. Nat Rev Immunol 2002;2:725–734. aspects. Subcell Biochem 2012;56:1–19. 20. Atreya I, Atreya R, Neurath MF. NF-kappaB in inflam- 3. Madsen CT, Sylvestersen KB, Young C, Larsen SC, matory bowel disease. J Intern Med 2008;263:591–596. Poulsen JW, Andersen MA, Palmqvist EA, Hey- 21. Singh S, Feuerstein JD, Binion DG, Tremaine WJ. AGA Mogensen M, Jensen PB, Treebak JT, Lisby M, technical review on the management of mild-to- Nielsen ML. Biotin starvation causes mitochondrial pro- moderate ulcerative colitis. Gastroenterology 2019; tein hyperacetylation and partial rescue by the SIRT3-like 156:769–808 e29. deacetylase Hst4p. Nat Commun 2015;6:7726. 22. Dassopoulos T, Sultan S, Falck-Ytter YT, Inadomi JM, 4. Hassan YI, Zempleni J. Epigenetic regulation of chro- Hanauer SB. American Gastroenterological Association matin structure and gene function by biotin. J Nutr 2006; Institute technical review on the use of thiopurines, 136:1763–1765. methotrexate, and anti-TNF-alpha biologic drugs for the 5. Rodriguez-Melendez R, Zempleni J. Regulation of gene induction and maintenance of remission in inflammatory expression by biotin (review). J Nutr Biochem 2003; Crohn’s disease. Gastroenterology 2013; 14:680–690. 145:1464–1478, e1-5. 6. Agrawal S, Agrawal A, Said HM. Biotin deficiency en- 23. Rubin DT, Ananthakrishnan AN, Siegel CA, Sauer BG, hances the inflammatory response of human dendritic Long MD. ACG clinical guideline: ulcerative colitis in cells. Am J Physiol Cell Physiol 2016;311:C386–C391. adults. Am J Gastroenterol 2019;114:384–413. 7. Baez-Saldana A, Diaz G, Espinoza B, Ortega E. Biotin 24. Neurath MF. Current and emerging therapeutic targets for deficiency induces changes in subpopulations of spleen IBD. Nat Rev Gastroenterol Hepatol 2017;14:269–278. lymphocytes in mice. Am J Clin Nutr 1998;67:431–437. 25. Antonelli E, Villanacci V, Bassotti G. Novel oral-targeted 8. Elahi A, Sabui S, Narasappa NN, Agrawal S, Lambrecht NW, therapies for mucosal healing in ulcerative colitis. World Agrawal A, Said HM. Biotin deficiency induces Th1- and J Gastroenterol 2018;24:5322–5330. Th17-mediated proinflammatory responses in human 26. Beaugerie L, Kirchgesner J. Balancing benefit vs risk of CD4(þ) T lymphocytes via activation of the mTOR signaling immunosuppressive therapy for individual patients with pathway. J Immunol 2018;200:2563–2570. inflammatory bowel diseases. Clin Gastroenterol Hepatol 9. Kuroishi T. Regulation of immunological and inflamma- 2019;17:370–379. tory functions by biotin. Can J Physiol Pharmacol 2015; 27. Fernandez-Banares F, Abad-Lacruz A, Xiol X, Gine JJ, 93:1091–1096. Dolz C, Cabre E, Esteve M, Gonzalez-Huix F, 10. Kuroishi T, Endo Y, Muramoto K, Sugawara S. Biotin Gassull MA. Vitamin status in patients with inflammatory deficiency up-regulates TNF-alpha production in murine bowel disease. Am J Gastroenterol 1989;84:744–748. macrophages. J Leukoc Biol 2008;83:912–920. 28. Abad-Lacruz A, Fernandez-Banares F, Cabre E, Gil A, 11. Subramanian VS, Constantinescu AR, Benke PJ, Esteve M, Gonzalez-Huix F, Xiol X, Gassull MA. The ef- Said HM. Mutations in SLC5A6 associated with brain, fect of total enteral tube feeding on the vitamin status of immune, bone, and intestinal dysfunction in a young malnourished patients with inflammatory bowel disease. child. Hum Genet 2017;136:253–261. Int J Vitam Nutr Res 1988;58:428–435. 12. Trueb RM. Serum biotin levels in women complaining of 29. Okabe N, Urabe K, Fujita K, Yamamoto T, Yao T, Doi S. hair loss. Int J Trichology 2016;8:73–77. Biotin effects in Crohn’s disease. Dig Dis Sci 1988; – 13. Ghosal A, Lambrecht N, Subramanya SB, Kapadia R, 33:1495 1496. Said HM. Conditional knockout of the Slc5a6 gene in 30. Urabe K. [Decreased plasma biotin levels in patients with mouse intestine impairs biotin absorption. Am J Physiol Crohn’s disease]. Nihon Shokakibyo Gakkai Zasshi Gastrointest Liver Physiol 2013;304:G64–G71. 1986;83:697. 2020 Biotin Supplementation Ameliorates Colitis 567

31. Vahid F, Rashvand S, Sadeghi M, Hekmatdoost A. The mechanisms. Am J Physiol Gastrointest Liver Physiol association between index of nutritional quality and ul- 2014;307:G941–G949. cerative colitis: a case-control study. J Res Med Sci 48. Ghosal A, Sekar TV, Said HM. Biotin uptake by mouse 2018;23:67. and human pancreatic beta cells/islets: a regulated, 32. Edgar R, Domrachev M, Lash AE. Gene Expression lipopolysaccharide-sensitive carrier-mediated process. Omnibus: NCBI gene expression and hybridization array Am J Physiol Gastrointest Liver Physiol 2014; data repository. Nucleic Acids Res 2002;30:207–210. 307:G365–G373. 33. Noble CL, Abbas AR, Cornelius J, Lees CW, Ho GT, 49. Subramanya SB, Subramanian VS, Kumar JS, Toy K, Modrusan Z, Pal N, Zhong F, Chalasani S, Clark H, Hoiness R, Said HM. Inhibition of intestinal biotin ab- Arnott ID, Penman ID, Satsangi J, Diehl L. Regional vari- sorption by chronic alcohol feeding: cellular and molec- ation in gene expression in the healthy colon is dysregu- ular mechanisms. Am J Physiol Gastrointest Liver lated in ulcerative colitis. Gut 2008;57:1398–1405. Physiol 2011;300:G494–G501. 34. Sabui S, Bohl JA, Kapadia R, Cogburn K, Ghosal A, 50. Cominelli F, Arseneau KO, Rodriguez-Palacios A, Lambrecht NW, Said HM. Role of the sodium-dependent Pizarro TT. Uncovering pathogenic mechanisms of in- multivitamin transporter (SMVT) in the maintenance of flammatory bowel disease using mouse models of intestinal mucosal integrity. Am J Physiol Gastrointest Crohn’s disease-like ileitis: what is the right model? Cell Liver Physiol 2016;311:G561–G570. Mol Gastroenterol Hepatol 2017;4:19–32. 35. Sabui S, Kapadia R, Ghosal A, Schneider M, 51. Kiesler P, Fuss IJ, Strober W. Experimental models of Lambrecht NWG, Said HM. Biotin and pantothenic acid inflammatory bowel diseases. Cell Mol Gastroenterol oversupplementation to conditional SLC5A6 KO mice Hepatol 2015;1:154–170. prevents the development of intestinal mucosal abnor- 52. Said HM, Mock DM, Collins JC. Regulation of intestinal malities and growth defects. Am J Physiol Cell Physiol biotin transport in the rat: effect of biotin deficiency 2018;315:C73–C79. and supplementation. Am J Physiol 1989; 36. Jeengar MK, Thummuri D, Magnusson M, Naidu VGM, 256:G306–G311. Uppugunduri S. Uridine ameliorates dextran sulfate so- 53. Sha T, Igaki K, Yamasaki M, Watanabe T, Tsuchimori N. dium (DSS)-induced colitis in mice. Sci Rep 2017;7:3924. Establishment and validation of a new semi-chronic 37. Papadakis KA, Targan SR. Role of cytokines in the dextran sulfate sodium-induced model of colitis in pathogenesis of inflammatory bowel disease. Annu Rev mice. Int Immunopharmacol 2013;15:23–29. Med 2000;51:289–298. 54. Murthy S, Murthy NS, Coppola D, Wood DL. The efficacy 38. de Souza HS, Fiocchi C. Immunopathogenesis of IBD: of BAY y 1015 in dextran sulfate model of mouse colitis. current state of the art. Nat Rev Gastroenterol Hepatol Inflamm Res 1997;46:224–233. 2016;13:13–27. 55. Gupta J, Nebreda AR. Analysis of intestinal permeability 39. Ghosal A, Jellbauer S, Kapadia R, Raffatellu M, Said HM. in mice. Bioprotocol 2014;4:e1289. Salmonella infection inhibits intestinal biotin transport: 56. Viennois E, Tahsin A, Merlin D. Purification of total RNA cellular and molecular mechanisms. Am J Physiol Gas- from DSS-treated murine tissue via lithium chloride pre- trointest Liver Physiol 2015;309:G123–G131. cipitation. Bio Protoc 2018;8. 40. Camilleri M. Leaky gut: mechanisms, measurement and clinical implications in humans. Gut 2019;68:1516–1526. Received June 28, 2019. Accepted November 21, 2019. 41. Mankertz J, Schulzke JD. Altered permeability in in- flammatory bowel disease: pathophysiology and clinical Correspondence implications. Curr Opin Gastroenterol 2007;23:379–383. Address correspondence to: Jonathan Skupsky, MD, PhD, Department of Medicine, Gastroenterology, University of California Irvine, 285 Irvine Hall, 42. Ma TY, Nighot P, Al-Sadi R. Tight junctions and the in- Irvine, California 92697. e-mail: [email protected]; fax: (949) 824-8540. testinal barrier. In: Said HM, ed. Physiology of the Acknowledgment gastrointestinal tract. London, England: Academic Press, The authors thank David Merriott for medical chart review. 2018:587–639. Author contributions 43. Giridharan S, Srinivasan M. Mechanisms of NF-kappaB Jonathan Skupsky, Subrata Sabui, and Hamid M. Said were responsible for the p65 and strategies for therapeutic manipulation. study concept and design; Jonathan Skupsky, Subrata Sabui, and Michael JInflamm Res 2018;11:407–419. Hwang acquired data; Jonathan Skupsky, Subrata Sabui, and Manando Nakasaki analyzed and interpreted the data; Jonathan Skupsky drafted the 44. Oeckinghaus A, Ghosh S. The NF-kappaB family of manuscript; Michael D. Cahalan and Hamid M. Said critically revised the transcription factors and its regulation. Cold Spring Harb manuscript; Jonathan Skupsky and Subrata Sabui performed the statistical analysis; Jonathan Skupsky, Michael D. Cahalan, and Hamid M. Said Perspect Biol 2009;1:a000034. obtained funding; Jonathan Skupsky, Subrata Sabui, and Michael Hwang 45. Bowman BB, Rosenberg IH. Biotin absorption by distal provided technical support; and Michael D. Cahalan and Hamid M. Said rat intestine. J Nutr 1987;117:2121–2126. supervised the study. 46. Mock D. Biotin. In: Rucker RB, Zempleni J, Suttie JW, Conflicts of interest McCormick DB, eds. Handbook of vitamins. 4th ed. The authors disclose no conflicts. Boca Raton: CRC Press, 2007:361–377. Funding 47. Srinivasan P, Kapadia R, Biswas A, Said HM. Chronic This work was supported by the Veteran’s Administration grants 5IK2BX003518 (J.S.) and I01BX001142 (H.M.S.), and by the National alcohol exposure inhibits biotin uptake by pancreatic Institutes of Health grants TR001415 (J.S.), NS14609 (M.D.C.), DK58057 acinar cells: possible involvement of epigenetic (H.M.S.), DK56061 (H.M.S.), and AA018071 (H.M.S.).