R E S EARCH ART I C L E

MICROB IOTA

Vitamin B12 modulates the transcriptome of the skinmicrobiota in acne pathogenesisDezhi Kang,1 Baochen Shi,1 Marie C. Erfe,2 Noah Craft,2 Huiying Li1,3*

http://stmD

ownloaded from

Various diseases have been linked to the human microbiota, but the underlying molecular mechanisms of themicrobiota in disease pathogenesis are often poorly understood. Using acne as a disease model, we aimed tounderstand the molecular response of the skin microbiota to host metabolite signaling in disease patho-genesis. Metatranscriptomic analysis revealed that the transcriptional profiles of the skin microbiota separatedacne patients from healthy individuals. The vitamin B12 biosynthesis pathway in the skin bacterium Propioni-bacterium acnes was significantly down-regulated in acne patients. We hypothesized that host vitamin B12modulates the activities of the skin microbiota and contributes to acne pathogenesis. To test this hypothesis,we analyzed the skin microbiota in healthy subjects supplemented with vitamin B12. We found that the supple-mentation repressed the expression of vitamin B12 biosynthesis genes in P. acnes and altered the transcriptomeof the skin microbiota. One of the 10 subjects studied developed acne 1 week after vitamin B12 supplemen-tation. To further understand the molecular mechanism, we revealed that vitamin B12 supplementation in P. acnescultures promoted the production of porphyrins, which have been shown to induce inflammation in acne. Ourfindings suggest a new bacterial pathogenesis pathway in acne and provide one molecular explanation for thelong-standing clinical observation that vitamin B12 supplementation leads to acne development in a subset of in-dividuals. Our study discovered that vitamin B12, an essential nutrient in humans, modulates the transcriptionalactivities of skin bacteria, and provided evidence that metabolite-mediated interactions between the host andthe skin microbiota play essential roles in disease development.

.scie

by guest on January 22, 2020

ncemag.org/

INTRODUCTION

Microorganisms live symbiotically with humans and contribute to hu-man health and disease. The taxonomic composition of the humanmicrobiome has been characterized in both healthy and diseased pop-ulations (1–3), revealing that alterations in the microbiome compositionare associated with certain diseases (4, 5). However, the underlying mo-lecular mechanisms of the human microbiota in disease pathogenesisare not well understood. At many body sites, including the skin, it hasnot been described whether host metabolites modulate the transcriptionalactivities of the microbiota. It is unclear whether the microbiota respondsto the host signals via altered metabolic activities, thus affecting thehost health/disease state. Understanding the molecular mechanisms ofhost-microbiota interactions may lead to more targeted therapy of manymicrobe-related human diseases.

Acne vulgaris (commonly called acne) provides a promising diseasemodel to study the interactions between the host and the skin micro-biota in disease pathogenesis because the microbiota is less complexand the disease has been associated with a single dominant bacterium,Propionibacterium acnes (2, 6–8). As one of the most common skindiseases, acne affects more than 80% of adolescents and young adults glob-ally (9, 10). Although not fatal, it can be extremely painful, disfiguring,and scarring, and in many patients, it can profoundly affect self-esteemand mental health (11, 12). Acne is a disease of the pilosebaceous unit(hair follicle), a unique skin compartment where resident bacteria in-teract with the host cells. Four main factors are believed to contribute

1Department of Molecular and Medical Pharmacology, Crump Institute for MolecularImaging, David Geffen School of Medicine, University of California, Los Angeles(UCLA), Los Angeles, CA 90095, USA. 2Los Angeles Biomedical Research Institute atHarbor–UCLA Medical Center, Torrance, CA 90502, USA. 3UCLA–Department ofEnergy (DOE) Institute for Genomics and Proteomics, Los Angeles, CA 90095, USA.*Corresponding author. E-mail: huiying@mednet.ucla.edu

www.Scien

to acne development: increased sebum production, follicular hyper-keratinization, proliferation of skin bacteria, and inflammation (13–16).The bacterial pathogenesis mechanism of acne remains to be defined.P. acnes has been long thought as a pathogenic factor for acne. However,it is a major skin commensal and dominates the skin microbiota in bothacne patients and healthy individuals (2). Understanding whether thetranscriptional activities of the skin microbiota are different betweenthe two cohorts would provide key insights on the bacterial pathogenesisof acne.

RESULTS

The transcriptional activities of the skin microbiota in acnepatients were distinct from those in healthy individualsTo determine whether the transcriptional activities of the skin micro-biota contribute to disease development, we compared the metatran-scriptome of the skin microbiota in acne patients to the one in healthyindividuals in a cross-sectional study. We collected the follicular con-tents of the nose skin from four acne patients and five healthy indi-viduals (Supplementary Materials). We analyzed the microbial geneexpression using RNA sequencing (RNA-Seq) with a high sequenc-ing depth of 44 to 182 million paired-end reads per sample (3.7 to18.1 giga base pairs) (table S1). We found that P. acneswas the most tran-scriptionally abundant bacterium (fig. S1). Additionally, Staphylococcus,Pseudomonas, and Shigella were detected in the metatranscriptome withmuch lower total transcriptional abundance. Given the predominanceof P. acnes in all samples, we focused our further analysis on the geneexpression profile of P. acnes.

We first quantified the P. acnes gene expression levels. Because eachindividual harbors various strains of P. acnes and the strain population

ceTranslationalMedicine.org 24 June 2015 Vol 7 Issue 293 293ra103 1

R E S EARCH ART I C L E

bhttp://stm

.sciencemag.org/

Dow

nloaded from

structures of P. acnes are different among individuals (2), we created anonredundant P. acnes gene set representative of all the genes encodedin various P. acnes strains. This enabled us to compare the gene ex-pression level among different individuals even though they may har-bor different strains. We binned the orthologous genes in 71 P. acnesgenomes, which cover all the major lineages of P. acnes found onhuman skin (17), into 5140 operational gene units (OGUs) as de-scribed in Materials and Methods. We then determined the expressionlevel of each OGU in each sample. We identified a core set of 3725 P.acnes OGUs (72.5% of all OGUs) expressed in all samples. Thiscore set of OGUs covered most of the metabolic pathways encodedin P. acnes genomes, including carbohydrate metabolism, nucleic acidmetabolism, amino acid metabolism, lipid metabolism, and the me-tabolism of cofactors and vitamins (fig. S2).

We next determined whether the gene expression profiles of P. acneswere distinct between acne patients and healthy individuals. We groupedthe samples on the basis of the expression profiles of P. acnes OGUsusing an unsupervised hierarchical clustering algorithm. Microbiomesamples of the acne patients formed a separate cluster from the samples ofthe healthy individuals (bootstrap value = 99%) (Fig. 1A). We furthershowed that acne patients were significantly overrepresented in thecluster (P = 0.0079, Fisher’s exact test). Other factors including gender,age, ethnicity, skin type, and hormone regulation were not significant-ly associated with the separation of the acne patients from the healthyindividuals. These results demonstrate that P. acnes transcriptional ac-tivities in the skin microbiota of acne patients were distinct from thoseof healthy individuals. To our knowledge, it is a pioneering findingthat the in vivo transcriptional activities of the microbiota clearly sep-arate the healthy and disease states of the host. This finding also high-lights the importance of studying the metatranscriptome of the humanmicrobiota in addition to the microbial composition to better un-

www.Scien

derstand the role of the microbial community in human health anddisease.

P. acnes vitamin B12 biosynthesis pathway wasdown-regulated in acne patientsTo determine the molecular mechanism of the skin microbiota in ac-ne pathogenesis, we identified the microbial genes and pathways thatwere differentially expressed between acne patients and healthy indi-viduals. In acne patients, 109 P. acnes OGUs were up-regulated and27 OGUs were down-regulated (Fig. 1B). Many of these differentiallyexpressed OGUs encode cellular components that are involved in me-tabolite and protein transport, including metal transporters (iron, co-balt, and hemin), multidrug transporters, protein export systems, andtype II bacterial secretion systems. A few previously proposed bacte-rial factors involved in acne pathogenesis (18) were found to be up-regulated in acne patients, including a putative lipase (PPA1035), aputative adhesion protein (PPA1907), and a conserved hypotheticalprotein that may be involved in capsular polysaccharide biosynthesis(PPA0150).

Using a functional annotation clustering analysis, DAVID (Databasefor Annotation, Visualization, and Integrated Discovery) (19), we iden-tified several differentially expressed metabolic pathways in acne. Theyinclude vitamin B12 biosynthesis and porphyrin metabolism, proteolysis,transport, arginine metabolic process, and glutamine family amino acidmetabolism. In parallel, using the algorithm of ShotgunFunctionalizeR,we again identified that the vitamin B12 biosynthesis pathway was sig-nificantly down-regulated (P = 9.3 × 10−51, ShotgunFunctionalizeR).We further mapped the differentially expressed OGUs to KEGG (KyotoEncyclopedia of Genes and Genomes) pathways (Fig. 2 and fig. S3).It indicated that in acne patients, P. acnes had increased transcrip-tional activities of the genes that encode carbohydrate uptake systems

ceTranslationalMedicine.org

y guest on January 22, 2020

and enzymes that catalyze carbohydratemetabolism, and decreased transcriptionalactivities of the genes that encode en-zymes in fatty acid biosynthesis and vita-min B12 biosynthesis.

Vitamin B12 has been implicated inacne pathogenesis. A number of clinicalstudies reported that supplementation ofvitamin B12 induced acne in a subset ofindividuals (20–26). However, the molec-ular mechanism of vitamin B12 in acnedevelopment is not understood. To bet-ter understand the role of vitamin B12 inacne pathogenesis, we first verified thedown-regulation of vitamin B12 biosyn-thesis genes in the skin microbiota in anew cohort of subjects. We collected skinmicrobial samples from an additional9 acne patients and 15 healthy individuals.Using quantitative real-time fluorescencepolymerase chain reaction (qRT-PCR),we quantified the expression levels of threegenes in the vitamin B12 biosynthesis path-way. Among the three genes, cysG+cbiX (afusion gene) encodes cobalamin synthesisprotein, and cbiL encodes precorrin-2 C20-methyltransferase. The enzymes encoded

Fig. 1. The gene expression profiles of P. acnes in the skin microbiota were distinct between acnepatients and healthy individuals. (A) On the basis of the gene expression of P. acnes in the skin micro-

biota, acne patients (labeled in red, “Acne”) formed a separate cluster from healthy individuals (labeled ingreen, “Healthy”) in an unsupervised hierarchical clustering analysis. (B) One hundred thirty-six differen-tially expressed P. acnes OGUs were identified between acne patients and healthy individuals. Amongthem, 109 OGUs were up-regulated and 27 OGUs were down-regulated in acne patients. The OGU namesare listed in table S2.

24 June 2015 Vol 7 Issue 293 293ra103 2

R E S EARCH ART I C L E

by guest on January 22, 2020http://stm

.sciencemag.org/

Dow

nloaded from

by these two genes function at the initial steps of corrin ring synthesiscontrolling the metabolic flow entering the vitamin B12 biosynthesis path-way (27). The third gene, btuR, encodes cob(I)alamin adenosyltransferase,which adenosylates the synthesized corrin ring, and once inactivated,blocks the synthesis of vitamin B12 (27). Consistent with the RNA-Seqdata obtained from our previous cohort, cysG+cbiX fusion gene and cbiLwere significantly down-regulated in the nine acne patients (fold change,3.34 and 1.94; P = 0.0023 and 0.0049, respectively, Student’s t test; n = 9acne patients and 15 healthy individuals). Gene btuR also had lower ex-pression levels in the acne patients compared to the healthy individuals,

www.Scien

although the difference was not statistically significant (fold change, 1.35;P = 0.17, Student’s t test; n = 9 acne patients and 15 healthy individuals)(Fig. 3). In this new cohort of subjects, we verified that vitamin B12 bio-synthesis gene expression in P. acnes was down-regulated in acne pa-tients compared to healthy individuals.

Vitamin B12 supplementation in healthy subjects repressedvitamin B12 biosynthesis genes in P. acnesIt is known that vitamin B12 is a regulator of its own biosynthesis path-way. Previous studies in other bacterial species, including Salmonella

Fig. 2. A schematic of the metabolic pathways in P. acnes illustratingthe observed differentially expressed OGUs in acne patients com-

metabolism pathways were up-regulated in acne patients. Genes in thevitamin B12 biosynthesis and fatty acid biosynthesis pathways were down-

pared to healthy individuals. Genes encoding sugar transport and sugar

regulated. FC, fold change. The full gene names are listed in table S3.

ceTranslationalMedicine.org 24 June 2015 Vol 7 Issue 293 293ra103 3

R E S EARCH ART I C L E

by guest on January 22, 2020http://stm

.sciencemag.org/

Dow

nloaded from

typhimurium (28) and Escherichia coli (29), showed that vitamin B12repressed the expression of cob/cbi operons in the vitamin B12 bio-synthesis pathway through cobalamin riboswitches. Cobalamin ribos-witches are conserved RNA structural elements that regulate theexpression of vitamin B12 biosynthesis operons upon vitamin B12binding (30). In P. acnes, the upstream regions of cob/cbi operons alsoencode cobalamin riboswitches (fig. S4).

To determine whether vitamin B12 represses P. acnes vitamin B12biosynthesis pathway in vivo, we conducted a longitudinal study, inwhich the gene expression of the skin microbiota in healthy subjectswith vitamin B12 supplementation was analyzed. We sampled 10 healthysubjects with clear skin who were receiving intramuscular injectionof vitamin B12 (1-mg hydroxocobalamin) for general well-being. It isdocumented that a single intramuscular injection of 1-mg hydroxo-cobalamin leads to a severalfold increase of the serum vitamin B12level to the range of 1500 to 57,000 pg/ml, which lasts for at least2 weeks (31). The serum vitamin B12 level usually returns to normalrange (200 to 900 pg/ml) in 4 weeks after injection (31–33). Stankleralso showed that the skin vitamin B12 levels correlate with the serumvitamin B12 levels (34). We collected the follicular content of the noseskin from the 10 healthy subjects at three time points: immediatelybefore vitamin B12 injection as a baseline control, 2 days after the in-jection, and 14 days after the injection. We measured the gene expressionlevels of cysG+cbiX, cbiL, and btuR in the vitamin B12 biosynthesis pathwayusing qRT-PCR (Fig. 4). Compared to the expression levels before or2 days after vitamin B12 injection, these genes were significantly down-regulated on day 14 after the vitamin B12 injection (cysG+cbiX P valuesrange from 1.6 × 10−4 to 2.1 × 10−4, cbiL P values range from 7.1 × 10−4

to 7.7 × 10−4, and btuR P values range from 9.6 × 10−5 to 0.0025, Student’st test; n = 10). Moreover, the gene expression levels on day 14 weresimilar to those observed in the acne patients (Fig. 4). To control foreffects not related to vitamin B12 supplementation, we resampled 9 ofthe 10 subjects with healthy skin 3 months later using the same sam-pling procedure without vitamin B12 supplementation. For the healthysubjects without vitamin B12 supplementation, there were no signifi-

www.ScienceTranslationalMedicine.org

cant changes in the expression levels ofvitamin B12 biosynthesis genes betweenday 0 and day 14. There were also no signif-icant differences in the gene expressionlevels of the samples collected on day 0 be-tween the original sampling period and theresampling period 3 months later (Fig. 4).Our study demonstrates that vitamin B12supplementation in the host repressed thevitamin B12 biosynthesis genes in P. acnes.

To determine whether vitamin B12 reg-ulates the transcription of additional genesand pathways in skin bacteria, we analyzedthe transcriptome of the skin microbiota inhealthy subjects with vitamin B12 supple-mentation.We performed RNA-Seq to an-alyze P. acnes gene expression profiles ofthe samples collected on day 0 before sup-plementation and on day 14 after sup-plementation. We were able to constructsequencing libraries with good quality from7 samples collected on day 0 and 10 sam-ples collected on day 14. From the RNA-Seq

data, we identified 93 OGUs that were differentially expressed be-tween day 0 and day 14 (Fig. 5A). Although individual variations existin some of these OGUs, the overall trend in up- or down-regulation ofthe OGUs was consistent among individuals. In addition to the vita-min B12 biosynthesis genes, the expression of other metabolic geneswas also altered after vitamin B12 supplementation. PPA0750, whichencodes S-adenosyl-methyltransferase MraW, was up-regulated. Theactivity of this enzyme depends on S-adenosylmethionine, which re-quires vitamin B12 for its regeneration. The changes in gene expressionof the vitamin B12–dependent biological processes validated that ourdata reflected the regulatory effects of vitamin B12 on the transcriptionalactivities of P. acnes. Additionally, the gene expression of a number oftranscriptional regulatory components and DNA replication machin-ery components was altered after vitamin B12 supplementation. The ex-pression of transporter genes was also affected, including up-regulatedexpression of iron compound, sugar, and oligopeptide transporters anddown-regulated expression of drug, proline/glycine betaine, and hypo-xanthine/guanine transporters. A sialidase gene (PPA1560), which hasbeen suggested to function in sebocyte adherence and host tissue deg-radation (18), was up-regulated after vitamin B12 supplementation. Ourresults suggest that vitamin B12 supplementation to the host affects notonly the expression of vitamin B12 biosynthesis genes in P. acnes butalso the gene expression of other metabolic enzymes and virulencefactor.

One of the 10 subjects developed acne 1 week aftervitamin B12 supplementationConsistent with multiple clinical studies previously reporting thatvitamin B12 supplementation induces acne in subsets of individuals(20–26), here, 1 of the 10 healthy subjects, HL414, had multiple ery-thematous papules and comedones developed on the face 1 week aftervitamin B12 supplementation (Supplementary Materials). It is not under-stood how vitamin B12 supplementation leads to acne in some individualsand whether the vitamin B12–associated acne has a separate pathogenesismechanism. To determine whether vitamin B12 supplementation in

Fig. 3. Vitamin B12 biosynthesis genes were down-regulated in the skin microbiota of acne pa-tients compared to healthy individuals. Down-regulation of the genes in P. acnes vitamin B bio-

12

synthesis pathway was validated in a separate cohort of acne patients (n = 9) and healthy individuals(n = 15). Consistent with the RNA-Seq data, cysG+cbiX and cbiL were significantly down-regulated, andbtuR showed a lower average expression level in acne patients compared to healthy individuals. Sig-nificance was determined by Student’s t test. The mean expression level of each gene is indicated by ablack bar. The gene expression data are listed in table S4.

24 June 2015 Vol 7 Issue 293 293ra103 4

R E S EARCH ART I C L E

by guest on January 22, 2020http://stm

.sciencemag.org/

Dow

nloaded from

subject HL414 altered the gene expression of the skin microbiota witha profile similar to those observed in the acne patients, we clusteredthe P. acnes gene expression profiles of the samples collected from

www.Scien

HL414 and the other healthy subjects with vitamin B12 supplementa-tion together with the samples from our earlier cross-sectional study(Fig. 5B). On day 0, the P. acnes gene expression pattern of HL414 wassimilar to other healthy subjects, but on day 14, it resembled the ex-pression pattern seen in the acne patients and was clustered within theacne group (bootstrap value = 100%). This suggests that the transcrip-tional changes in the skin microbiota of HL414 after vitamin B12 sup-plementation are not different from those in acne patients. This resultstrongly supports the hypothesis that the host vitamin B12 level mod-ulates the transcriptional activities of the skin bacteria, which in turnaffect the health or disease state of the host skin. The day 14 samplesfrom the nine subjects who did not develop acne after supplementa-tion were clustered outside of the acne group. Some of these samples,including HL417_Day14, HL431_Day14, and HL430_Day14, wereclustered with their corresponding day 0 samples. This finding sug-gests that the effects of the vitamin B12 supplementation on regulatingthe transcriptional activities of the skin microbiota vary among indi-viduals, possibly because of both host and microbial factors. These dif-ferences may partly explain why only a subset of individuals developsacne after vitamin B12 supplementation.

The transcriptional profile of the skin microbiotain HL414 was different from other healthy subjectsafter vitamin B12 supplementationTo understand why subject HL414, but not others in our healthy cohort,developed acne after vitamin B12 supplementation, we further investi-gated the specific bacterial transcriptional changes in HL414 comparedto the other subjects 14 days after vitamin B12 supplementation. On thebasis of the RNA-Seq data, we found 397 OGUs differentially expressedin HL414_Day14 compared to the other nine day 14 samples. We com-pared the list of these genes (HL414 versus others on day 14) with thegenes that were significantly altered after vitamin B12 supplementationin all healthy subjects (day 0 versus day 14 samples). This comparison isbased on the reasoning that bacterial factors involved in vitamin B12–associated acne should be regulated by vitamin B12. We found that 11OGUs that were differentially expressed between HL414_Day14 andthe other day 14 samples were also regulated by vitamin B12 with dif-ferential expression between day 0 and day 14 samples (table S6).Among them, PPA0693 encodes the E2 component of the 2-oxoglutaratedehydrogenase complex, which functions in the tricarboxylic acid (TCA)cycle. It was down-regulated in all the healthy subjects after vitamin B12supplementation (fold change, 0.86; P = 0.044, Student’s t test; n = 7).Among all the day 14 samples, HL414_Day14 had the lowest expression.This gene was also down-regulated when we compared the acne patientsto the healthy individuals in our cross-sectional study (fold change, 0.76;P = 0.075, Student’s t test; n = 4 acne patients and 5 healthy individuals)(fig. S5). This finding suggests that the gene down-regulation of the E2component of the 2-oxoglutarate dehydrogenase complex is regulated byvitamin B12 and is involved in acne pathogenesis.

We further investigated the potential role of PPA0693 in acne path-ogenesis. On the basis of the metabolic pathways in P. acnes (Fig. 2),2-oxoglutarate dehydrogenase complex converts 2-oxoglutarate to succinyl–coenzyme A (CoA). 2-Oxoglutarate is a substrate for the biosynthesis ofL-glutamate, which is a precursor for both vitamin B12 and porphyrinbiosynthesis (35). It has been shown that the biosynthesis of porphyrinsin Propionibacteria is inversely correlated with the biosynthesis of vita-min B12 (36, 37). Bykhovskiı̆ et al. (36) and Zaı̆tseva et al. (38) foundthat inhibiting vitamin B12 biosynthesis in Propionibacterium shermanii,

Fig. 4. Vitamin B12 supplementation in healthy subjects repressedP. acnes cob/cbi operons. The gene expression levels of cbiL, cysG+cbiX,

and btuR in P. acnes vitamin B12 biosynthesis pathway were significantlyrepressed in the healthy subjects (n = 10) on day 14 after vitamin B12 sup-plementation, to a level similar to those observed in the acne patients (n = 9).Without vitamin B12 supplementation, the expression levels of these genesdid not change significantly on day 2 and day 14 compared to day 0 (n = 9).The expression levels, quantified by qRT-PCR, for each gene on day 0, day 2,and day 14 are shown. Data from healthy subjects with vitamin B12 supple-mentation are shown in purple, and data from those without supplementa-tion are shown in green. As a comparison, data from the acne patients areshown in red. The mean expression level of each gene is indicated by a blackbar. Significance was determined by Student’s t test. The gene expressiondata are listed in tables S5A and S5B.

ceTranslationalMedicine.org 24 June 2015 Vol 7 Issue 293 293ra103 5

Fig. 5. Vitamin B12 supplementation in the host altered the transcrip-tome of P. acnes in the skin microbiota. (A) The differentially expressedP. acnes OGUs between day 0 and day 14 samples from the healthy sub-jects supplemented with vitamin B12. The red and green colors represent thefold change of the OGUs. Red indicates up-regulation after vitamin B12 sup-plementation and green indicates down-regulation. The functional categoriesof the OGUs are labeled by the colored bars next to the OGU names. (B)

The P. acnes gene expression profile of subject HL414 on day 14 after vita-min B12 supplementation (HL414_Day14) was similar to those from the acnepatients and clustered together. The day 0 sample from subject HL414(“HL414_Day0”, healthy skin, before vitamin B12 supplementation) was simi-lar to the samples from other healthy subjects. Samples from the acne pa-tients were labeled in red (“Acne”), and samples from the healthy subjectswere labeled in green (“Healthy,” “Day0,” “Day14”), except for HL414_Day14.

R E S EARCH ART I C L E

www.ScienceTranslationalMedicine.org 24 June 2015 Vol 7 Issue 293 293ra103 6

by guest on January 22, 2020http://stm

.sciencemag.org/

Dow

nloaded from

R E S EARCH ART I C L E

by guest on January 22, 2020http://stm

.sciencemag.org/

Dow

nloaded from

either by enzyme inhibitor or by depleting cobalt, led to an increase inporphyrin biosynthesis. Porphyrins produced by P. acnes are known toinduce inflammation in acne (39–41). They interact with molecular ox-ygen, generate free radicals to damage adjacent keratinocytes, and stim-ulate the production of inflammatory mediators in keratinocytes (39–41).

On the basis of the above data, we propose a molecular mechanismfor the bacterial pathogenesis of acne through vitamin B12 regulation.We hypothesize that elevated vitamin B12 level in the host modulatesthe transcriptional activities of the skin microbiota. Among the altera-tions in P. acnes, PPA0693 is down-regulated, resulting in an increasein 2-oxoglutarate and L-glutamate. The vitamin B12 biosynthesis path-way is also down-regulated, resulting in the increased metabolic flowof L-glutamate being shunted to the porphyrin biosynthesis pathway.These changes result in an overproduction of porphyrins by P. acnes,which can induce an inflammatory response in the host cells leading toacne. This hypothesis can potentially explain the molecular mechanismof vitamin B12 in acne pathogenesis. In a subset of individuals, similar toHL414, the expression level of PPA0693 after vitamin B12 supplemen-tation is lower than in other individuals, resulting in further increasedlevel of porphyrins produced and leading to acne development.

Porphyrin production is promoted in P. acnes withvitamin B12 supplementationTo test the above hypothesis, we investigated whether porphyrin pro-duction is promoted in P. acnes by vitamin B12 supplementation. We sup-plemented vitamin B12 (10 mg/ml) to P. acnes cultures and found thatvitamin B12 persistently repressed the expression of cob/cbi operons.The previously mentioned three genes in the vitamin B12 biosynthesispathway, cysG+cbiX, cbiL, and btuR, were significantly down-regulatedby the addition of vitamin B12 from day 2 to day 14 (P values range from0.057 to 0.0031, Student’s t test; n = 2 to 3) in cultures (Fig. 6A). In themeantime, we compared the amounts of porphyrins produced in theP. acnes cultures with or without vitamin B12 supplementation. We foundthat the addition of vitamin B12 significantly increased porphyrin produc-tion by 39% during the exponential phase on day 8 and nearly 19% atthe stationary phase on day 14 (P = 0.0043 and 6.8 × 10−4 respectively,Student’s t test; n = 3 to 4) (Fig. 6B). Our results demonstrated thatvitamin B12 supplementation repressed the gene expression of vitaminB12 biosynthesis in P. acnes and increased the biosynthesis of porphyrins.

To further determine that the increased porphyrin production is di-rectly linked to the reduction in vitamin B12 biosynthesis, we measuredthe porphyrin level in P. acnes under the condition that vitamin B12 bio-synthesis was repressed without the regulation by vitamin B12. We re-pressed vitamin B12 biosynthesis by depleting cobalt (Co

2+ ion), which isan essential component required for vitamin B12 biosynthesis (38). Weobserved that the porphyrin production in P. acnes cultures was signif-icantly increased after Co2+ depletion (fig. S6). This observation is con-sistent with previous studies in other Propionibacteria (36, 37). Ourresult demonstrates that the vitamin B12 and porphyrin biosynthesispathways share the same pool of precursor substrates and are inverselycorrelated at the metabolic level, further supporting that down-regulationof the vitamin B12 biosynthesis pathway promotes porphyrin production.

DISCUSSION

Originally reported more than a half-century ago, it is observed thatvitamin B12 leads to acne development in a subset of population. How-

www.Scien

ever, the underlying mechanism of this clinical observation was notunderstood. Here, by investigating the transcriptional activities andregulations of the skin microbiota, we revealed a molecular link be-tween vitamin B12 and acne pathogenesis. We found that the vitaminB12 biosynthesis pathway was down-regulated in acne patients and inhealthy subjects who were supplemented with vitamin B12 (Figs. 3 and4). Several lines of evidence suggest that a high level of serum vitaminB12 is associated with acne phenotype. Goldblatt et al. (42) found thatthe serum level of vitamin B12 was significantly higher in acne patientsthan in healthy individuals. Moreover, they and Karadag et al. reportedthat in acne patients, the serum level of vitamin B12 was significantlydecreased after treatment (42, 43). On the basis of our findings andthe evidence mentioned above, we propose a vitamin B12–mediatedbacterial mechanism for acne pathogenesis (Fig. 7). In healthy skin,when vitamin B12 level is normal, the vitamin B12 biosynthesis pathwayin the skin bacterium P. acnes is expressed and porphyrins are producedat a low level. When the host vitamin B12 level is elevated, it signals

Fig. 6. Vitamin B12 supplementation repressed the expression of cob/cbioperons in the vitamin B biosynthesis pathway and promoted

12

porphyrin production in P. acnes cultures. (A) The expression levels ofcbiL, cysG+cbiX, and btuR were significantly repressed on day 2, day 8, andday 14 after vitamin B12 supplementation in P. acnes cultures. The experi-ments were repeated three times independently. Each dot represents theaverage gene expression level of two to three technical replicates in eachindependent experiment. The dotted lines indicate the changes betweenthe P. acnes cultures in medium alone (black) and the cultures in mediumwith vitamin B12 supplementation (red). The mean of the gene expressionlevels from the three independent experiments is shown as a black bar.Significance was determined by Student’s t test. The gene expression dataare listed in table S7. (B) Porphyrin production in P. acnes cultures on day 8and day 14 after vitamin B12 supplementation was significantly increased.Significance was determined by Student’s t test. Data are presented asmeans ± SD. The experiments were repeated three times independentlywith three to four biological replicates each time. The porphyrin levelsare listed in table S8.

ceTranslationalMedicine.org 24 June 2015 Vol 7 Issue 293 293ra103 7

R E S EARCH ART I C L E

by guest on January 22, 2020http://stm

.sciencemag.org/

Dow

nloaded from

transcriptional changes in P. acnes. The levels of 2-oxoglutarate andL-glutamate are increased, and the vitamin B12 biosynthesis in P. acnesis repressed. The metabolic flow of L-glutamate is shunted toward theporphyrin biosynthesis pathway, leading to an overproduction of por-phyrins by P. acnes in the follicle. The overproduced porphyrins, se-creted by P. acnes, induce an inflammatory response in the host cells,leading to acne development. Supporting this proposed mechanism,clinical improvement after acne treatment has been associated witha lower level of porphyrins produced in follicles (44–47). Borelli et al.(46) also reported that in patients with poor treatment outcome, theporphyrin level was not decreased. Our proposed mechanism of vi-tamin B12–mediated interactions between the host and the skin mi-crobiota provides one potential molecular explanation for the bacterialpathogenesis of acne. It also emphasizes the importance of metabolite-mediated interactions between the host and the microbiota in humanhealth and disease.

Here, acne developed in subject HL414 in a short period after vi-tamin B12 supplementation, but not in other healthy subjects. Thisobservation is consistent with previous clinical findings that not allindividuals developed acne after vitamin B12 supplementation (20–26).We also found that after vitamin B12 supplementation, HL414 had

www.Scien

a significantly lower gene expression level in the E2 componentof the 2-oxoglutarate dehydrogenase complex compared to the oth-er subjects. The lower expression level of this gene may contribute to ahigher level of 2-oxoglutarate and L-glutamate in P. acnes, leading to agreater amount of porphyrins produced. This may partially explain whyHL414 developed acne whereas others did not during the 14-day pe-riod. In addition to PPA0693, we found several other bacterial factorsthat were associated with acne development after vitamin B12 supple-mentation (table S6). Among them, PPA2220 was up-regulated in allthe acne patients compared to the healthy individuals and in the healthysubjects after vitamin B12 supplementation (fig. S5). Moreover, subjectHL414 had a significantly higher expression level of PPA2220 than otherhealthy subjects after vitamin B12 supplementation. The function ofPPA2220 is not yet known. It encodes a protein containing a photo-synthetic reaction center barrel domain and a domain of unknown function(DUF2382). Its orthologs in other organisms are often located next tothe genes that encode electron transfer chain components, acyl-CoAtransferase, and/or dehydrogenases. PPA2220 was also the second mosthighly expressed gene in the P. acnes transcriptome in our samples.These findings suggest that PPA2220 may play an important role inacne pathogenesis and needs to be further studied. Other pathogenicfactors from the host and/or the bacteria, regulated by vitamin B12 ornot, may also be important in the disease development and need to beinvestigated in future studies.

To our knowledge, the metatranscriptome of the human skin mi-crobiota in health and disease states has not been described before thisstudy. Whereas some other human body sites, such as the oral cavity,gut, urogenital tract, and upper respiratory tract, can have sufficientbiomass for microbiota characterization (48–51), the skin presents achallenging site with low microbial biomass density (6, 52). Studies ofthe skin microbiome have been limited to mostly taxonomic composi-tion analysis (53). It is important to understand the transcriptional changesof the skin microbiota during diseases in addition to its taxonomic com-position and metagenomic content. Here, we applied RNA-Seq to quan-titatively measure the metatranscriptome of the skin microbiota in acnepatients and in healthy individuals. We revealed microbial differencesat the transcriptional level separating acne patients and healthy indi-viduals. By vitamin B12 supplementation in healthy individuals andin vitro experiments, we revealed that host vitamin B12 modulates thetranscriptional and metabolic activities of skin bacteria, leading toincreased production of porphyrins, which induce inflammation in ac-ne. Our findings suggest a new metabolite-mediated bacterial mecha-nism for acne pathogenesis.

Metabolite-mediated interactions between the host and the micro-biota play essential roles in human health and disease. This has beenrecognized in the gut microbiota (54). Host nutritional states or inter-ventions, such as diets and use of antibiotics and other small-moleculedrugs, have been shown to modulate the composition and transcrip-tional activities of the gut microbiota (55–58). In response, the gut mi-crobiota produces metabolites that are linked to diseases (54, 59–61).Vitamin B12 is an essential nutrient to humans. Its microbial biosynthesisand human absorption mainly take place in the gut. It has been shownthat vitamin B12 regulates gut microbial gene expression and affectsthe selection and competition among microbial species (62). Ourstudy showed that vitamin B12 supplementation to the host modulatesthe transcriptome of the skin microbiota and leads to acne develop-ment in a subset of individuals. Our study not only provides an ex-planation for the long-standing clinical observations of the correlation

Fig. 7. A model of vitamin B12 modulating the transcriptional and meta-bolic activities of the skin bacterium P. acnes in acne pathogenesis. In

healthy skin, when the host vitamin B12 level is normal, the vitamin B12 bio-synthesis pathway in P. acnes is expressed and porphyrins are produced ata low level. When the host vitamin B12 level is elevated, it signals transcrip-tional changes in P. acnes. The levels of 2-oxoglutarate and L-glutamateare increased, and the vitamin B12 biosynthesis in P. acnes is repressed. Themetabolic flow of L-glutamate is shunted toward the porphyrin biosynthesispathway, leading to an overproduction of porphyrins by P. acnes in the fol-licle. The overproduced porphyrins, secreted by P. acnes, induce an inflam-matory response in the host cells, leading to acne development in a subset ofindividuals. 2-OG, 2-oxoglutarate; L-Glu, L-glutamate.

ceTranslationalMedicine.org 24 June 2015 Vol 7 Issue 293 293ra103 8

R E S EARCH ART I C L E

Dow

nloaded

between increased vitamin B12 level and acne development but alsosuggests a molecular mechanism for acne pathogenesis. Future follow-up studies may potentially lead to the development of new therapeuticsfor this medically important disease.

There are a few limitations of our study. First, we had a limitednumber of subjects investigated in our longitudinal study. Amongthe 10 healthy subjects, one individual developed acne after vitaminB12 supplementation. Although it is consistent with the previous re-ports that vitamin B12 leads to acne development in only a subset ofindividuals (20–26), increasing the cohort size would render morepower in statistical analysis and would allow us to further establish themolecular link between vitamin B12 and acne pathogenesis. Second,the number of samples in our RNA-Seq metatranscriptomic analysis waslimited. This was mainly because skin samples have low bacterial biomassand it is difficult to extract sufficient amount of RNAs from each sam-ple. Some of our samples failed during sample processing, RNA ex-traction, and sequencing library construction. Future improvements ofsample collection and bacterial RNA extraction from samples withlimited biomass will empower us to better understand the microbialtranscriptional activities and their regulations on the human skin.

by guest on January 22, 2020http://stm

.sciencemag.org/

from

MATERIALS AND METHODS

Study designA cross-sectional study to compare the transcriptional activities of theskin microbiota between acne patients and healthy individuals wasperformed. Among the nine subjects recruited, four were acne patients(two males and two females) and five were healthy individuals (two malesand three females). The average age of the acne patients was 24.8 years(range, 19 to 38 years) and the average age of the healthy subjects was22.8 years (range, 13 to 32 years). Among the acne patients, two wereAfrican American and two were Hispanic. Among the healthy subjects,one was African American, one was Hispanic, two were Caucasian, andone was Asian. There were no significant differences in gender, age, andethnicity between the acne patients and the healthy subjects. The averageface acne score of the four acne patients was 2.5 and the average noseacne score was 0.25 (Supplementary Materials). Twenty-four additionalsubjects were recruited for the gene expression analysis of the vitaminB12 biosynthesis pathway. Among them, 9 were acne patients (threemales and six females) and 15 were healthy subjects (seven males andeight females). The average age of the acne patients was 23.8 years (range,15 to 42 years) and the average age of the healthy subjects was 33.8 years(range, 21 to 44 years). Among the acne patients, four were Hispanic,two were Caucasian, two were Asian, and one was of more than one race.Their average face acne score was 2.8, and the average nose acne scorewas 1.6 (Supplementary Materials). Among the healthy subjects, threewere Hispanic, three were Caucasian, eight were Asian, and one was ofmore than one race. There were no significant differences in gender andethnicity, but age, between the acne patients and the healthy subjects.

Among the 15 healthy subjects, 10 were receiving intramuscularinjection of vitamin B12 [1 ml of hydroxocobalamin (1000 mg/ml)] forgeneral well-being and were included in the longitudinal study. The lon-gitudinal study was performed to investigate the effects of vitamin B12supplementation on the transcriptional activity of the skin microbiotain healthy individuals.

None of the healthy subjects reported any current or past acne treat-ment. None of the acne patients were being treated with antibiotics at

www.Scien

the time of sampling. Four acne patients were being treated with othertopical acne therapies at the time of sampling, including tazarotene,tretinoin, salicylic acid, and benzoyl peroxide. Seven acne patients hadbeen treated in the past. None of the subjects had cosmetic procedures.

All subjects were recruited in Southern California. The diagnosis ofacne was made by board-certified dermatologists. The presence of acnewas graded on a scale of 0 to 5 relating closely to the Global AcneSeverity Scale (63). Subjects with healthy skin were determined byboard-certified dermatologists and were defined as people who hadno acneiform lesions on the face, chest, or back. All subjects providedwritten informed consent. All protocols and consent forms were ap-proved by the Institutional Review Boards of the University of California,Los Angeles (UCLA) and the Los Angeles Biomedical Research Insti-tute. The study was not blinded.

Sample collectionThe follicular contents of the nose skin were sampled from the sub-jects using Bioré Deep Cleansing Pore Strips (Kao Brands Company)following the instruction of the manufacturer (2). This sampling meth-od is different from the conventional tape stripping method. It samplesmostly the follicular content of the pilosebaceous unit, whereas the con-ventional tape stripping method samples the stratum corneum of theepidermis. A comparison of the samples collected from opposite sidesof the nose of the same individual using these two different methods isillustrated in fig. S7. Clean gloves were used for each sampling. Afterremoval from the nose, the strip was placed into a 50-ml sterile tubeand labeled with a coded sample name.

For the 10 healthy subjects with vitamin B12 supplementation, thefirst samples (day 0) were taken from one side of the nose before theyreceived vitamin B12 injection. The second samples (day 2) were takenfrom the other side of the nose 2 days after the injection. The third sam-ples (day 14) were taken from the entire nose 14 days after the injection.As a control, 3 months later, 9 of the 10 healthy subjects were sampledagain using the same sampling protocol over a period of 14 days with-out vitamin B12 supplementation.

Sequence analysisSequence reads mapped to the human genome were removed follow-ing the procedure used in the Human Microbiome Project (64). Low-quality reads were trimmed or filtered out (details in the SupplementaryMaterials). The ribosomal RNA (rRNA) reads were then removed bymapping against P. acnes 16S, 23S, and 5S rRNA sequences and againstthe SILVA database (release 108) (65). Sequence mapping was per-formed using Bowtie (version 0.12.7), allowing up to three mismatchesper read (66). The remaining reads were aligned to 71 P. acnes genomes(17). For paired-end mapping, we required that the read pairs werealigned to the same reference genome on different strands, with the dis-tance between the two reads shorter than 1 kb. The remaining readswere then mapped against the Human Microbiome Project reference ge-nomes (www.hmpdacc.org/HMREFG/). The resulting unaligned readswere further searched against the National Center for BiotechnologyInformation (NCBI) nonredundant nucleotide database (RefSeq re-lease 48), including bacterial, fungal, and viral genomes.

OGU constructionP. acnes OGUs were constructed in a similar way as previously described(67) with minor modification. Genes in all 71 P. acnes genomes werebinned on the basis of their protein sequences using CD-HIT (Cluster

ceTranslationalMedicine.org 24 June 2015 Vol 7 Issue 293 293ra103 9

R E S EARCH ART I C L E

by guest on January 22, 2020http://stm

.sciencemag.org/

Dow

nloaded from

Database at High Identity with Tolerance) version 4.3 (68) and its de-fault parameters (≥90% protein sequence identity).

Quantification of OGU expression levelWe assigned a coverage score of 1 to each reference nucleotide alignedby a read with unique alignment. The total coverage score of eachreference nucleotide was calculated by summing the scores of the nu-cleotide from all the aligned reads. The reads with multiple alignmenthits were analyzed as previously described (69). The coverage score ofeach nucleotide within the aligned region in the reference genome wasaveraged by the total number of hits (M).

The expression level of a gene was calculated as the total coveragescore for all the nucleotides within the gene region normalized by thegene length. To adjust for sequencing depth differences among dif-ferent samples, the expression level of each P. acnes gene was furthernormalized by the total number of base pairs aligned to P. acnes gen-omes. Reads aligned to P. acnes rRNAs and transfer-messenger RNAswere removed before normalization. OGU expression level was calcu-lated by summing the normalized expression level of each gene mem-ber of the OGU.

Porphyrin quantification in P. acnes culturesP. acnes was cultured in reinforced clostridial broth anaerobically withoutexposure to light at 37°C. The culture medium was supplemented withvitamin B12 (Sigma) at 10 mg/ml, a concentration similar to the oneused in a previous study with E. coli (70). The controls were culturedwithout vitamin B12 supplementation. After 2 days, 8 days, and 14 days,200 ml of bacterial culture was used to measure optical density at 595nm, and 500 ml of bacterial culture was used to extract porphyrins usingthe method described previously (71) with minor modifications. Briefly,bacterial culture was mixed with 250 ml of ethyl acetate/acetic acid(4:1) for 10 s by vortexing and was then centrifuged for 5 min at12,000 rpm. The upper phase was transferred to a new tube and mixedwith 250 ml 1.5 M HCl for 10 s. After centrifugation for 2 min at 12,000rpm, 200 ml of extracted porphyrins in the HCl lower phase was takenand quantified by measuring its absorbance at 405 nm. All the opticalabsorbance was quantified on a GENios microplate reader (Tecan). Theabsorbance at 405 nm was then converted to concentration based on astandard curve determined using coproporphyrin III.

cob/cbi gene expression analysisTotal RNA was extracted from P. acnes cell cultures or skin microbialsamples collected from the subjects using the protocol described in theSupplementary Materials. The total RNA was converted to single-stranded complementary DNA using SuperScript III First-Strand Syn-thesis SuperMix (Life Technologies). qRT-PCR was performed on aLightCycler 480 (Roche) using the LightCycler 480 High ResolutionMelting Master Mix (Roche) and the following primers: cbiL-forward,5′-GCGCGAGGCAGACGTGATCC-3′; cbiL-reverse, 5′-GACACCG-GACCTCTCCCGCA-3′; cysG+cbiX-forward, 5′-TGTATTCCGCC-CCGCTGTTGC-3′; cysG+cbiX-reverse, 5′-GAGCACTGCCGACGTGTCCC-3′;btuR-forward, 5′-GGAAGATGCTCTTCGGGCGCT-3′; btuR-reverse,5′-GCCTCAGGGTTCTCCGCAGC-3′; 16S-forward, 5′-GGGGCTT-AACCCTGAGCGTGC-3′; 16S-reverse, 5′-TTCGCTCCCCACGCTT-TCGC-3′. The qRT-PCR protocol was set as follows: initial denaturationat 95°C for 5 min, followed by 50 cycles of 95°C for 10 s, 62°C for 30 s,and 72°C for 30 s. The expression level of each gene was expressed asthe logarithm of its relative expression level to 16S rRNA transcript level.

www.Scien

StatisticsThe differentially expressed OGUs in the RNA-Seq data were identi-fied using Student’s t test (details in the Supplementary Materials). Thethreshold of statistical significance was set at P < 0.05. The differentiallyexpressed OGUs were further confirmed by ShotgunFunctionalizeR(72), with a cutoff of Akaike’s information criterion < 5000 and adjustedP < 0.05. The differentially expressed KEGG pathways in acne patientswere determined by ShotgunFunctionalizeR, with a cutoff of Akaike’sinformation criterion < 5000 and adjusted P < 0.05. DAVID (19) wasused to identify the enriched metabolic pathways and functional anno-tation clusters. Student’s t test was used to determine the statistical sig-nificance of the differences in vitamin B12 gene expression quantified byqRT-PCR and in porphyrin production between the culture conditions.Unsupervised hierarchical clustering was generated on the basis of thecorrelation metric of OGU expression levels, and the average linkageclustering method was used. The robustness of the clustering was eval-uated by the bootstrapping method using the R package pvclust (73).The significance of the association of the skin clinical status, as wellas other subject information, with the clustering was determined byFisher’s exact test.

SUPPLEMENTARY MATERIALS

www.sciencetranslationalmedicine.org/cgi/content/full/7/293/293ra103/DC1Materials and MethodsFig. S1. The metatranscriptome composition of the skin microbiota in follicles.Fig. S2. The KEGG pathways expressed in all the samples.Fig. S3. The KEGG pathways with differentially expressed P. acnes OGUs.Fig. S4. The cob/cbi operons in P. acnes.Fig. S5. Differentially expressed P. acnes OGUs in sample HL414_Day14 compared to the otherday 14 samples.Fig. S6. The biosynthesis of porphyrins is inversely correlated with the biosynthesis of vitaminB12 in P. acnes.Fig. S7. A comparison of the samples collected from opposite sides of the nose of the sameindividual using the Bioré strip method and the tape stripping method.Fig. S8. Rarefaction curves indicate sufficient sequencing depths of the samples.Table S1. High sequencing depths of the metatranscriptomic data.Table S2. Differentially expressed OGUs shown in Fig. 1B.Table S3. Full names of the genes and substrates shown in Fig. 2.Table S4. The relative expression levels (log10) of the vitamin B12 biosynthesis genes in the skinmicrobiota of the healthy individuals and acne patients shown in Fig. 3.Table S5A. The relative expression levels (log10) of the vitamin B12 biosynthesis genes in theskin microbiota of the healthy subjects shown in Fig. 4.Table S5B. The relative expression levels (log10) of the vitamin B12 biosynthesis genes in theskin microbiota of the acne patients shown in Fig. 4.Table S6. The differentially expressed OGUs in both the comparison between sampleHL414_Day14 and other day 14 samples and the comparison between day 0 and day 14samples in the vitamin B12 supplementation study.Table S7. The relative expression levels (log10) of the vitamin B12 biosynthesis genes in the P. acnescultures shown in Fig. 6A.Table S8. The porphyrin levels detected in the P. acnes cultures shown in Fig. 6B.References (74–82)

REFERENCES AND NOTES

1. The Human Microbiome Project Consortium, Structure, function and diversity of thehealthy human microbiome. Nature 486, 207–214 (2012).

2. S. Fitz-Gibbon, S. Tomida, B.-H. Chiu, L. Nguyen, C. Du, M. Liu, D. Elashoff, M. C. Erfe, A. Loncaric,J. Kim, R. L. Modlin, J. F. Miller, E. Sodergren, N. Craft, G. M. Weinstock, H. Li, Propionibacteriumacnes strain populations in the human skin microbiome associated with acne. J. Invest.Dermatol. 133, 2152–2160 (2013).

ceTranslationalMedicine.org 24 June 2015 Vol 7 Issue 293 293ra103 10

R E S EARCH ART I C L E

by guest on January 22, 2020http://stm

.sciencemag.org/

Dow

nloaded from

3. E. A. Grice, H. H. Kong, S. Conlan, C. B. Deming, J. Davis, A. C. Young; NISC ComparativeSequencing Program, G. G. Bouffard, R. W. Blakesley, P. R. Murray, E. D. Green, M. L. Turner,J. A. Segre, Topographical and temporal diversity of the human skin microbiome. Science324, 1190–1192 (2009).

4. I. Cho, M. J. Blaser, The human microbiome: At the interface of health and disease. Nat.Rev. Genet. 13, 260–270 (2012).

5. K. J. Pflughoeft, J. Versalovic, Human microbiome in health and disease. Annu. Rev. Pathol.7, 99–122 (2012).

6. R. R. Marples, J. J. Leyden, R. N. Stewart, O. H. Mills Jr., A. M. Kligman, The skin microflora inacne vulgaris. J. Invest. Dermatol. 62, 37–41 (1974).

7. J. J. Leyden, K. J. McGinley, O. H. Mills, A. M. Kligman, Propionibacterium levels in patientswith and without acne vulgaris. J. Invest. Dermatol. 65, 382–384 (1975).

8. K. T. Holland, W. J. Cunliffe, C. D. Roberts, The role of bacteria in acne vulgaris: A newapproach. Clin. Exp. Dermatol. 3, 253–257 (1978).

9. W. J. Cunliffe, D. J. Gould, Prevalence of facial acne vulgaris in late adolescence and inadults. Br. Med. J. 1, 1109–1110 (1979).

10. S. Z. Ghodsi, H. Orawa, C. C. Zouboulis, Prevalence, severity, and severity risk factors ofacne in high school pupils: A community-based study. J. Invest. Dermatol. 129, 2136–2141 (2009).

11. M. A. Gupta, A. K. Gupta, Depression and suicidal ideation in dermatology patients withacne, alopecia areata, atopic dermatitis and psoriasis. Br. J. Dermatol. 139, 846–850(1998).

12. J. A. Halvorsen, R. S. Stern, F. Dalgard, M. Thoresen, E. Bjertness, L. Lien, Suicidal ideation,mental health problems, and social impairment are increased in adolescents with acne: Apopulation-based study. J. Invest. Dermatol. 131, 363–370 (2011).

13. H. Gollnick, Current concepts of the pathogenesis of acne: Implications for drug treatment.Drugs 63, 1579–1596 (2003).

14. U. Jappe, Pathological mechanisms of acne with special emphasis on Propionibacteriumacnes and related therapy. Acta Derm. Venereol. 83, 241–248 (2003).

15. G. F. Webster, Inflammation in acne vulgaris. J. Am. Acad. Dermatol. 33, 247–253 (1995).16. H. C. Williams, R. P. Dellavalle, S. Garner, Acne vulgaris. Lancet 379, 361–372 (2012).17. S. Tomida, L. Nguyen, B.-H. Chiu, J. Liu, E. Sodergren, G. M. Weinstock, H. Li, Pan-genome

and comparative genome analyses of Propionibacterium acnes reveal its genomic diversityin the healthy and diseased human skin microbiome. MBio 4, e00003–e00013 (2013).

18. H. Brüggemann, A. Henne, F. Hoster, H. Liesegang, A. Wiezer, A. Strittmatter, S. Hujer, P. Dürre,G. Gottschalk, The complete genome sequence of Propionibacterium acnes, a commensal ofhuman skin. Science 305, 671–673 (2004).

19. G. Dennis Jr., B. T. Sherman, D. A. Hosack, J. Yang, W. Gao, H. C. Lane, R. A. Lempicki, DAVID:Database for Annotation, Visualization, and Integrated Discovery. Genome Biol. 4, P3(2003).

20. W. Jadassohn, R. Paillard, R. Hofer, M. Golaz, Vitamine B12 et poussée aenéïforme. Dermatologica116, 349 (1958).

21. A. Puissant, F. Vanbremeersch, J. Monfort, J. N. Lamberton, A new iatrogenic dermatosis:Acne caused by vitamin B 12. Bull. Soc. Fr. Dermatol. Syphiligr. 74, 813–815 (1967).

22. P. Dugois, P. Amblard, R. Imbert, B. de Bignicourt, Acne caused by vitamin B 12. Lyon Med.221, 1165 (1969).

23. O. Braun-Falco, H. Lincke, The problem of vitamin B6/B12 acne. A contribution on acnemedicamentosa (author’s transl). MMW Munch. Med. Wochenschr. 118, 155–160 (1976).

24. A. Dupré, N. Albarel, J. Bonafe, B. Christol, J. Lassere, Vitamin B-12 induced acnes. Cutis 24,210 (1979).

25. E. Sherertz, Acneiform eruption due to “megadose” vitamins B6 and B12. Cutis 48, 119–120(1991).

26. I. Balta, P. Ozuguz, Vitamin B12-induced acneiform eruption. Cutan. Ocul. Toxicol. 33, 94–95(2014).

27. M. J. Warren, E. Raux, H. L. Schubert, J. C. Escalante-Semerena, The biosynthesis of adeno-sylcobalamin (vitamin B12). Nat. Prod. Rep. 19, 390–412 (2002).

28. S. Ravnum, D. I. Andersson, Vitamin B12 repression of the btuB gene in Salmonella typhimuriumis mediated via a translational control which requires leader and coding sequences. Mol.Microbiol. 23, 35–42 (1997).

29. X. Nou, R. J. Kadner, Coupled changes in translation and transcription during cobalamin-dependent regulation of btuB expression in Escherichia coli. J. Bacteriol. 180, 6719–6728(1998).

30. A. G. Vitreschak, D. A. Rodionov, A. A. Mironov, M. S. Gelfand, Regulation of the vitamin B12metabolism and transport in bacteria by a conserved RNA structural element. RNA 9,1084–1097 (2003).

31. G. B. Glass, H. R. Skeggs, D. H. Lee, E. L. Jones, W. W. Hardy, Hydroxocobalamin. I. Bloodlevels and urinary excretion of vitamin B12 in man after a single parenteral dose of aqueoushydroxocobalamin, aqueous cyanocobalamin and cyanocobalamin zinc-tannate complex.Blood 18, 511–521 (1961).

32. S. A. Tauber, H. R. Skeggs, A. Itkin, J. L. Stanley, Parenteral studies with vitamin B12 complexes.Am. J. Clin. Nutr. 10, 480–483 (1962).

www.Scien

33. A. P. Skouby, Retention and distribution of B12 activity, and requirement for B12, followingparenteral administration of hydroxocobalamin (Vibeden). Acta Med. Scand. 180, 95–105(1966).

34. L. Stankler, The vitamin B12 level in psoriatic skin and serum. Br. J. Dermatol. 81, 911–918(1969).

35. K. Murakami, Y. Hashimoto, Y. Murooka, Cloning and characterization of the gene encod-ing glutamate 1-semialdehyde 2,1-aminomutase, which is involved in d-aminolevulinicacid synthesis in Propionibacterium freudenreichii. Appl. Environ. Microbiol. 59, 347–350 (1993).

36. V. Bykhovskiĭ, N. I. Zaĭtseva, V. N. Bukin, Possible competition in the use of delta-aminolevulinicacid for the biosynthesis of vitamin B 12 and porphyrin by resting suspensions of Propioni-bacterium shermanii. Dokl. Akad. Nauk SSSR 180, 232–234 (1968).

37. H. C. Friedmann, L. M. Cagen, Microbial biosynthesis of B12-like compounds. Annu. Rev.Microbiol. 24, 159–208 (1970).

38. N. I. Zaĭtseva, V. Bykhovskiĭ, V. N. Bukin, Regulation of vitamin B-12 and prophyrin biosynthesisin Propionibacterium shermanii. Dokl. Akad. Nauk SSSR 190, 1476–1479 (1970).

39. D. Saint-Leger, A. Bague, E. Cohen, M. Chivot, A possible role for squalene in the pathoge-nesis of acne. I. In vitro study of squalene oxidation. Br. J. Dermatol. 114, 535–542 (1986).

40. E. M. Gribbon, J. G. Shoesmith, W. J. Cunliffe, K. T. Holland, The microaerophily and photo-sensitivity of Propionibacterium acnes. J. Appl. Bacteriol. 77, 583–590 (1994).

41. M. Schaller, M. Loewenstein, C. Borelli, K. Jacob, M. Vogeser, W. H. Burgdorf, G. Plewig, Induc-tion of a chemoattractive proinflammatory cytokine response after stimulation of keratino-cytes with Propionibacterium acnes and coproporphyrin III. Br. J. Dermatol. 153, 66–71 (2005).

42. S. Goldblatt, Vitamin B-12 levels in blood serum in acne-like skin diseases. Hautarzt 17,106–108 (1966).

43. A. S. Karadag, E. Tutal, D. T. Ertugrul, K. O. Akin, Effect of isotretinoin treatment on plasmaholotranscobalamin, vitamin B12, folic acid, and homocysteine levels: Non-controlledstudy. Int. J. Dermatol. 50, 1564–1569 (2011).

44. R. J. Martin, G. Kahn, J. W. Gooding, G. Brown, Cutaneous porphyrin fluorescence as anindicator of antibiotic absorption and effectiveness. Cutis 12, 758–764 (1973).

45. O. H. Mills Jr., A. M. Kligman, P. Pochi, H. Comite, Comparing 2.5%, 5%, and 10% benzoylperoxide on inflammatory acne vulgaris. Int. J. Dermatol. 25, 664–667 (1986).

46. C. Borelli, K. Merk, M. Schaller, K. Jacob, M. Vogeser, G. Weindl, U. Berger, G. Plewig, In vivoporphyrin production by P. acnes in untreated acne patients and its modulation by acnetreatment. Acta Derm. Venereol. 86, 316–319 (2006).

47. S. Mayr-Kanhäuser, B. Kränke, W. Aberer, Efficacy of octenidine dihydrochloride and2-phenoxyethanol in the topical treatment of inflammatory acne. Acta Dermatovenerol.Alp. Panonica Adriat. 17, 139–143 (2008).

48. G. L. Simon, S. L. Gorbach, Intestinal flora in health and disease. Gastroenterology 86, 174–193(1984).

49. J. Washio, T. Sato, T. Koseki, N. Takahashi, Hydrogen sulfide-producing bacteria in tonguebiofilm and their relationship with oral malodour. J. Med. Microbiol. 54, 889–895 (2005).

50. E. S. Charlson, K. Bittinger, A. R. Haas, A. S. Fitzgerald, I. Frank, A. Yadav, F. D. Bushman,R. G. Collman, Topographical continuity of bacterial populations in the healthy humanrespiratory tract. Am. J. Respir. Crit. Care Med. 184, 957–963 (2011).

51. Z. Ling, X. Liu, X. Chen, H. Zhu, K. E. Nelson, Y. Xia, L. Li, C. Xiang, Diversity of cervicovaginalmicrobiota associated with female lower genital tract infections. Microb. Ecol. 61, 704–714(2011).

52. J. J. Leyden, K. J. McGinley, K. M. Nordstrom, G. F. Webster, Skin microflora. J. Invest. Dermatol.88, 65s–72s (1987).

53. E. A. Grice, J. A. Segre, The skin microbiome. Nat. Rev. Microbiol. 9, 244–253 (2011).54. V. Tremaroli, F. Bäckhed, Functional interactions between the gut microbiota and host

metabolism. Nature 489, 242–249 (2012).55. I. Cho, S. Yamanishi, L. Cox, B. A. Methé, J. Zavadil, K. Li, Z. Gao, D. Mahana, K. Raju, I. Teitler,

H. Li, A. V. Alekseyenko, M. J. Blaser, Antibiotics in early life alter the murine colonic micro-biome and adiposity. Nature 488, 621–626 (2012).

56. L. A. David, C. F. Maurice, R. N. Carmody, D. B. Gootenberg, J. E. Button, B. E. Wolfe, A. V. Ling,A. S. Devlin, Y. Varma, M. A. Fischbach, S. B. Biddinger, R. J. Dutton, P. J. Turnbaugh, Dietrapidly and reproducibly alters the human gut microbiome. Nature 505, 559–563 (2014).

57. H. J. Haiser, D. B. Gootenberg, K. Chatman, G. Sirasani, E. P. Balskus, P. J. Turnbaugh, Predictingand manipulating cardiac drug inactivation by the human gut bacterium Eggerthella lenta.Science 341, 295–298 (2013).

58. G. D. Wu, J. Chen, C. Hoffmann, K. Bittinger, Y. Y. Chen, S. A. Keilbaugh, M. Bewtra, D. Knights,W. A. Walters, R. Knight, R. Sinha, E. Gilroy, K. Gupta, R. Baldassano, L. Nessel, H. Li, F. D. Bushman,J. D. Lewis, Linking long-term dietary patterns with gut microbial enterotypes. Science 334,105–108 (2011).

59. Y. Furusawa, Y. Obata, S. Fukuda, T. A. Endo, G. Nakato, D. Takahashi, Y. Nakanishi, C. Uetake,K. Kato, T. Kato, M. Takahashi, N. N. Fukuda, S. Murakami, E. Miyauchi, S. Hino, K. Atarashi,S. Onawa, Y. Fujimura, T. Lockett, J. M. Clarke, D. L. Topping, M. Tomita, S. Hori, O. Ohara,T. Morita, H. Koseki, J. Kikuchi, K. Honda, K. Hase, H. Ohno, Commensal microbe-derivedbutyrate induces the differentiation of colonic regulatory T cells. Nature 504, 446–450(2013).

ceTranslationalMedicine.org 24 June 2015 Vol 7 Issue 293 293ra103 11

R E S EARCH ART I C L E

by guest on January 2http://stm

.sciencemag.org/

Dow

nloaded from

60. Z. Wang, E. Klipfell, B. J. Bennett, R. Koeth, B. S. Levison, B. DuGar, A. E. Feldstein, E. B. Britt,X. Fu, Y.-M. Chung, Y. Wu, P. Schauer, J. D. Smith, H. Allayee, W. H. Tang, J. A. DiDonato, A. J. Lusis,S. L. Hazen, Gut flora metabolism of phosphatidylcholine promotes cardiovascular disease.Nature 472, 57–63 (2011).

61. S. Yoshimoto, T. M. Loo, K. Atarashi, H. Kanda, S. Sato, S. Oyadomari, Y. Iwakura, K. Oshima,H. Morita, M. Hattori, K. Honda, Y. Ishikawa, E. Hara, N. Ohtani, Obesity-induced gut microbialmetabolite promotes liver cancer through senescence secretome. Nature 499, 97–101(2013).

62. A. L. Goodman, N. P. McNulty, Y. Zhao, D. Leip, R. D. Mitra, C. A. Lozupone, R. Knight, J. I. Gordon,Identifying genetic determinants needed to establish a human gut symbiont in its habitat.Cell Host Microbe 6, 279–289 (2009).

63. B. Dréno, F. Poli, H. Pawin, C. Beylot, M. Faure, M. Chivot, N. Auffret, D. Moyse, F. Ballanger,J. Revuz, Development and evaluation of a Global Acne Severity Scale (GEA Scale) suitablefor France and Europe. J. Eur. Acad. Dermatol. Venereol. 25, 43–48 (2011).

64. The Human Microbiome Project Consortium, A framework for human microbiome re-search. Nature 486, 215–221 (2012).

65. C. Quast, E. Pruesse, P. Yilmaz, J. Gerken, T. Schweer, P. Yarza, J. Peplies, F. O. Glockner, TheSILVA ribosomal RNA gene database project: Improved data processing and web-based tools.Nucleic Acids Res. 41, D590–D596 (2013).

66. B. Langmead, C. Trapnell, M. Pop, S. L. Salzberg, Ultrafast and memory-efficient alignmentof short DNA sequences to the human genome. Genome Biol. 10, R25 (2009).

67. E. E. Hansen, C. A. Lozupone, F. E. Rey, M. Wu, J. L. Guruge, A. Narra, J. Goodfellow, J. R. Zaneveld,D. T. McDonald, J. A. Goodrich, A. C. Heath, R. Knight, J. I. Gordon, Pan-genome of the dom-inant human gut-associated archaeon, Methanobrevibacter smithii, studied in twins. Proc. Natl.Acad. Sci. U.S.A. 108, 4599–4606 (2011).

68. W. Li, L. Jaroszewski, A. Godzik, Clustering of highly homologous sequences to reduce thesize of large protein databases. Bioinformatics 17, 282–283 (2001).

69. C. M. Sharma, S. Hoffmann, F. Darfeuille, J. Reignier, S. Findeiß, A. Sittka, S. Chabas, K. Reiche,J. Hackermüller, R. Reinhardt, P. F. Stadler, J. Vogel, The primary transcriptome of the majorhuman pathogen Helicobacter pylori. Nature 464, 250–255 (2010).

70. M. D. Lundrigan, R. J. Kadner, Altered cobalamin metabolism in Escherichia coli btuR mu-tants affects btuB gene regulation. J. Bacteriol. 171, 154–161 (1989).

71. S. Piomelli, B. Davidow, V. F. Guinee, P. Young, G. Gay, The FEP (free erythrocyte porphyr-ins) test: A screening micromethod for lead poisoning. Pediatrics 51, 254–259 (1973).

72. E. Kristiansson, P. Hugenholtz, D. Dalevi, ShotgunFunctionalizeR: An R-package for func-tional comparison of metagenomes. Bioinformatics 25, 2737–2738 (2009).

73. R. Suzuki, H. Shimodaira, Pvclust: An R package for assessing the uncertainty in hierarchicalclustering. Bioinformatics 22, 1540–1542 (2006).

74. J. Frias-Lopez, Y. Shi, G. W. Tyson, M. L. Coleman, S. C. Schuster, S. W. Chisholm, E. F. Delong,Microbial community gene expression in ocean surface waters. Proc. Natl. Acad. Sci. U.S A. 105,3805–3810 (2008).

75. R. S. Poretsky, N. Bano, A. Buchan, G. LeCleir, J. Kleikemper, M. Pickering, W. M. Pate, M. A. Moran,J. T. Hollibaugh, Analysis of microbial gene transcripts in environmental samples. Appl. Environ.Microbiol. 71, 4121–4126 (2005).

76. R. L. Tatusov, N. D. Fedorova, J. D. Jackson, A. R. Jacobs, B. Kiryutin, E. V. Koonin, D. M. Krylov,R. Mazumder, S. L. Mekhedov, A. N. Nikolskaya, B. S. Rao, S. Smirnov, A. V. Sverdlov, S. Vasudevan,

www.Scien

Y. I. Wolf, J. J. Yin, D. A. Natale, The COG database: An updated version includes eukaryotes.BMC Bioinformatics 4, 41 (2003).

77. M. Kanehisa, S. Goto, Y. Sato, M. Furumichi, M. Tanabe, KEGG for integration and interpre-tation of large-scale molecular data sets. Nucleic Acids Res. 40, D109–D114 (2012).

78. J. Oksanen, F. G. Blanchet, R. Kindt, P. Legendre, R. G. O’Hara, G. L. Simpson, P. Solymos,M. Henry, H. Stevens, H. Wagner, vegan: Community ecology package. R package version1.17-0 (2010); http://cran.r-project.org/web/packages/vegan/.

79. B. Bolker, L. Bonebakker, R. Gentleman, W. H. A. Liaw, T. Lumley, M. Maechler, A. Magnusson,S. Moeller, M. Schwartz, B. Venables, gplots: Various R programming tools for plotting data. Rpackage version 2.7.4 (2009); http://cran.r-project.org/web/packages/gplots/.

80. A. S. Cristino, E. D. Tanaka, M. Rubio, M.-D. Piulachs, X. Belles, Deep sequencing of organ- andstage-specific microRNAs in the evolutionarily basal insect Blattella germanica (L.) (Dictyoptera,Blattellidae). PLOS One 6, e19350 (2011).

81. T. Yamada, I. Letunic, S. Okuda, M. Kanehisa, P. Bork, iPath2.0: Interactive pathway explorer.Nucleic Acids Res. 39, W412–W415 (2011).

82. G. Kasimatis, S. Fitz-Gibbon, S. Tomida, M. Wong, H. Li, Analysis of complete genomes ofPropionibacterium acnes reveals a novel plasmid and increased pseudogenes in an acneassociated strain. Biomed. Res. Int. 2013, 918320 (2013).

Acknowledgments:We thank S. Tomida for suggestions on experimental design and A.C. Maretti-Mirafor assistance in sample collection. We thank E. Barnard, T. Johnson, B. Chiu, J. Torres, and L. Altierifor assistance in performing experiments and data analysis. B. Tan provided the photographiccomparison of the two different skin sampling methods. We also thank E. Barnard, D. Eisenberg,Z. Guo, M. Liu, M. Pellegrini, and B. Tan for suggestions on improving the manuscript. The Uni-versity of California Davis Sequencing Center and UCLA Broad Stem Cell Research Center Sequen-cing Center provided sequencing support. Funding: This study was funded by NIH grantsR01GM099530 from the National Institute of General Medical Sciences and UH2AR057503 fromthe National Institute of Arthritis and Musculoskeletal and Skin Diseases. Author contributions:D.K. performed the experiments, analyzed the data, and wrote the manuscript. B.S. analyzed thedata and revised the manuscript. M.C.E. collected clinical samples. N.C. conceived and directedthe clinical study, recruited subjects, collected samples, interpreted clinical correlations and data,and revised the manuscript. H.L. conceived and directed the project, analyzed the data, andwrote the manuscript. Competing interests: The Regents of the University of California is theowner of a patent application (US Application No. 62/078,275, titled “Compositions andmethods for treating or preventing acne”) related to the role of vitamin B12 in acne, whichnames D.K., B.S., N.C., and H.L. as inventors. The other author declares no competing interests.Data and materials availability: The RNA-Seq reads have been deposited in the NCBISequence Read Archive (BioProject number PRJNA260091).

Submitted 25 March 2015Accepted 27 May 2015Published 24 June 201510.1126/scitranslmed.aab2009

Citation: D. Kang, B. Shi, M. C. Erfe, N. Craft, H. Li, Vitamin B12 modulates the transcriptome ofthe skin microbiota in acne pathogenesis. Sci. Transl. Med. 7, 293ra103 (2015).

2

ceTranslationalMedicine.org 24 June 2015 Vol 7 Issue 293 293ra103 12

, 2020

modulates the transcriptome of the skin microbiota in acne pathogenesis12Vitamin BDezhi Kang, Baochen Shi, Marie C. Erfe, Noah Craft and Huiying Li

DOI: 10.1126/scitranslmed.aab2009, 293ra103293ra103.7Sci Transl Med

changes in the skin microbiome and disease pathology.resulting in the production of inflammation-inducing porphyrins. These studies demonstrate a clear link between

and altered the transcriptome of the skin microbiota,Propionibacterium acnes synthesis genes in 12vitamin B reduced the expression of12induced acne. Supplementing patients with vitamin B−12contribute to vitamin B

. show that transcriptional changes in the skin microbiotaet aleffect has remained unexplored. Now, Kang supplements have been shown to lead to acne development, but the mechanism behind this12Vitamin B

contributes to acne by affecting the skin microbiota12Vitamin B

ARTICLE TOOLS http://stm.sciencemag.org/content/7/293/293ra103

MATERIALSSUPPLEMENTARY http://stm.sciencemag.org/content/suppl/2015/06/22/7.293.293ra103.DC1

CONTENTRELATED

http://stm.sciencemag.org/content/scitransmed/11/492/eaau8465.fullhttp://stm.sciencemag.org/content/scitransmed/11/490/eaat8329.fullhttp://science.sciencemag.org/content/sci/363/6424/227.fullhttp://science.sciencemag.org/content/sci/352/6285/539.fullhttp://science.sciencemag.org/content/sci/352/6285/532.fullhttp://science.sciencemag.org/content/sci/352/6285/544.fullhttp://science.sciencemag.org/content/sci/352/6285/535.fullhttp://science.sciencemag.org/content/sci/352/6285/560.fullhttp://stke.sciencemag.org/content/sigtrans/10/475/eaan4459.fullhttp://stm.sciencemag.org/content/scitransmed/9/378/eaah4680.fullhttp://science.sciencemag.org/content/sci/352/6285/565.fullhttp://science.sciencemag.org/content/sci/349/6253/1172.fullhttp://science.sciencemag.org/content/sci/352/6289/1116.fullhttp://science.sciencemag.org/content/sci/352/6285/586.fullhttp://science.sciencemag.org/content/sci/352/6285/530.fullhttp://science.sciencemag.org/content/sci/351/6279/1296.fullhttp://science.sciencemag.org/content/sci/350/6256/aac5992.fullhttp://stke.sciencemag.org/content/sigtrans/5/253/ec307.abstracthttp://stm.sciencemag.org/content/scitransmed/4/137/137rv5.fullhttp://stm.sciencemag.org/content/scitransmed/4/137/137rv7.full

REFERENCES

http://stm.sciencemag.org/content/7/293/293ra103#BIBLThis article cites 80 articles, 15 of which you can access for free

Terms of ServiceUse of this article is subject to the

registered trademark of AAAS. is aScience Translational MedicineScience, 1200 New York Avenue NW, Washington, DC 20005. The title

(ISSN 1946-6242) is published by the American Association for the Advancement ofScience Translational Medicine

Copyright © 2015, American Association for the Advancement of Science

by guest on January 22, 2020http://stm

.sciencemag.org/

Dow

nloaded from

PERMISSIONS http://www.sciencemag.org/help/reprints-and-permissions

Terms of ServiceUse of this article is subject to the

registered trademark of AAAS. is aScience Translational MedicineScience, 1200 New York Avenue NW, Washington, DC 20005. The title

(ISSN 1946-6242) is published by the American Association for the Advancement ofScience Translational Medicine

Copyright © 2015, American Association for the Advancement of Science

by guest on January 22, 2020http://stm

.sciencemag.org/

Dow

nloaded from