A-D, Facial telangiectasia and erythema tended to be located on the central face in ETR. E-H, Telangiectasia and erythema tended to be located on the lateral face in TP. A and E, Circles show biopsy locations; in both conditions, biopsies were centered on a dilated vessel. B-D and F-H, Enlargements show clinical features of the central face (B and F), preauricular area (C and G), and jawline (D and H).
The x-axis numbers 0 through 4 indicate degrees of severity on a scale explained in detail for each domain in eTable 1 in the Supplement. Briefly, 0 indicates a normal, healthy state in all domains; 4 indicates the most severe condition in the specified domain.
A, Comparison of characteristic light microscopic features (top; hematoxylin-eosin) and transmission electron micrographic features (bottom). Under light microscopy, the ETR specimen has dilated vessels surrounded by intact collagen (green asterisks) and a greater degree of inflammation than the TP specimen; TP specimen has dilated vessels surrounded by more fragmented collagen (black asterisks) and a greater degree of solar elastosis (yellow asterisks). Under electron microscopy, the ETR specimen shows dilated blood vessel with multiple intact adjacent collagen fibrils (inset shows crosscut collagen, original magnification ×400); the TP specimen shows dilated blood vessel surrounded by areas of unstructured, damaged fibers (inset, original magnification ×400) and some amorphous material consistent with elastotic material (yellow asterisk). B, Light microscopic features of mast cells (mast cell tryptase stain). The control specimen shows stain localized to mast cells (inset, original magnification ×400). The ETR specimen shows staining of mast cells in addition to light brown staining of the area surrounding mast cells, indicative of mast cell degranulation (inset, original magnification ×400). C, Immunohistochemical quantitative image analysis of mast cell tryptase staining; area demonstrated greater than 4-fold increased staining in ETR vs TP, and 25-fold greater staining in ETR vs control. There was 6-fold increased staining in TP vs control.
A-E, In all panels, asterisks indicate P < .05, and error bars, standard errors. A, For neuropeptide gene transcripts, RT-PCR showed induction of gene expression of calcitonin gene-related peptides (CGRP)-α (CALCA), CGRP-β (CALCB), and substance P (TAC1) in ETR compared with Ctrl, and upregulation of CGRP-α and substance P compared with TP. B, For collagen gene transcripts, RT-PCR showed upregulated gene expression of types I (COL-1) and III (COL-3) collagen in ETR compared with Ctrl and TP. C, For matrix remodeling gene transcripts decorin (DCN) and Cyr-61 (CYR61), RT-PCR revealed upregulated gene expression in ETR compared with Ctrl and TP. D, For matrix metalloproteinases (MMPs)-1, 3, and 9 gene transcripts, RT-PCR showed upregulation of all 3 MMPs in ETR compared with Ctrl; MMP-3 was also upregulated compared with TP; MMP-1 and MMP-9 were upregulated in TP compared with Ctrl. E, For gene transcripts involved in mast cell chemotaxis, RT-PCR showed upregulation of chemokine CXCL12 and its receptor, CXCR4, in ETR compared with Ctrl; CXCR4 was also upregulated in ETR compared with TP.
eTable 1. Histologic features examined
eTable 2. Custom PCR primers and probes
eTable 3. Fold induction of gene transcripts relevant to rosacea as determined by RT-PCR
Helfrich YR, Maier LE, Cui Y, Fisher GJ, Chubb H, Fligiel S, Sachs D, Varani J, Voorhees J. Clinical, Histologic, and Molecular Analysis of Differences Between Erythematotelangiectatic Rosacea and Telangiectatic Photoaging. JAMA Dermatol. 2015;151(8):825-836. doi:10.1001/jamadermatol.2014.4728
Facial erythema and telangiectasia are commonly associated with the erythematotelangiectatic subtype of rosacea (ETR). It is important for clinicians to recognize that these findings can also be associated with a subtype of photoaging, which we term telangiectatic photoaging (TP).
To demonstrate that ETR and TP are distinct dermatologic disorders.
A case-control observational study comparing clinical, histologic, and gene expression features of 26 participants with ETR, 20 with TP, and 11 age- and sex-matched controls in the Program for Clinical Research in Dermatology at University of Michigan.
Main Outcomes and Measures
Findings of clinical history and examination, light and electron microscopy, immunohistochemical analyses, and real-time quantitative reverse-transcriptase polymerase chain reaction gene expression.
Transient erythema was greater in the ETR group (38% graded moderate to severe) than in the TP (0%; P < .001) and control groups (0%; P = .002). Nontransient erythema was also greater in the ETR group (50% graded moderate to severe) than in the TP (25%; P = .03) and control groups (0%; P < .001). Participants with ETR tended to have erythema and telangiectasia primarily on the central face (79%), whereas those with TP tended to have more lateral involvement (57%; P < .001). Those with ETR had significantly less clinical evidence of photodamage (0% graded 6-8 on a photonumeric scale) than those with TP (40% graded 6-8; P = .01). Histologically, there was less evidence of photodamage in ETR than in TP, which had wispy collagen and solar elastosis surrounding blood vessels. Immunohistologic analysis demonstrated greater geometric mean immunostained area by mast cell tryptase staining in ETR samples (0.018%) than in TP (0.004%; P = .01) or control samples (0.001%; P < .001) but no increase in mast cell number, indicative of greater mast cell degranulation. Gene expression of matrix metalloproteinase-3 was 4-fold greater in ETR samples than in TP samples (P = .004) and 5-fold higher than in control samples (P = .004). Gene expression of the neuropeptides calcitonin gene-related peptide (CGRP-α) and substance P was significantly increased in ETR compared with TP (9-fold [P < .001] and 5-fold [P = .002], respectively) and control samples (10-fold [P < .001] and 28-fold [P < .001], respectively).
Conclusions and Relevance
Telangiectatic photoaging is characterized by less transient and nontransient erythema, a more lateral distribution of erythema and telangiectasia, less neurogenic mast cell activation, and less MMP-mediated matrix remodeling than ETR. These data demonstrate that TP is a distinct clinical entity from ETR that can be distinguished on the basis of clinical, histologic, and gene expression findings.
In 2002, the National Rosacea Society published a consensus statement outlining 4 rosacea subtypes—papulopustular, erythematotelangiectatic, phymatous, and ocular.1 Of these subtypes, erythematotelangiectatic rosacea (ETR) is probably the most disputed. Some clinicians argue that it is merely photodamage2; certainly it can be difficult to distinguish from actinic damage.3 The National Rosacea Society Expert Committee described fairly permissive criteria for defining rosacea. The presence of 1 or more of the following, typically involving the convexities of the face, is sufficient: (1) transient erythema (flushing); (2) nontransient erythema; (3) papules and pustules; and/or (4) telangiectasia. The guidelines state as follows: “Erythematotelangiectatic rosacea is mainly characterized by flushing and persistent central facial erythema. The appearance of telangiectases is common but not essential for a diagnosis of this subtype.”1(p585)
We and others have observed that significant telangiectasia and persistent facial erythema can be observed not only in patients with rosacea but also in patients with characteristic features of photodamage. Quiz Ref IDThese patients, typically older men, lack other rosacea symptoms such as flushing, burning, or stinging, and they often do not complain of facial redness, which is instead an incidental finding. Facial erythema and telangiectasia tend to be more lateral but can involve the central face and nose. Others have labeled this finding actinic telangiectasia.2 We call it telangiectatic photoaging (TP).
We assert that patients with TP are a distinct clinical population and do not have ETR. We set out to explore and better define the clinical, histologic, and molecular distinctions between ETR and TP, with an eventual goal of aiding physicians in diagnosing these conditions and appropriately tailoring patient treatments and expectations.
This case-control observational study was approved by the University of Michigan institutional review board, and all participants provided their written informed consent.
Participants with visible facial telangiectasia and erythema were recruited via local announcement and advertisement as well as from clinics within the University of Michigan Dermatology Department. All enrolled participants were willing to undergo facial biopsies and facial photography and met the criteria for diagnosis of either rosacea or TP. Participants underwent a comprehensive history and physical examination by a faculty member or a second- or third-year resident dermatologist trained in clinical research in our Program for Clinical Research in Dermatology. On the basis of the findings from this history and physical examination, participants were classified as having ETR or TP.
Participants defined as having ETR had transient erythema (flushing), nontransient erythema, and/or facial telangiectasia—criteria established in the National Rosacea Society consensus statement1; for the purpose of this study, telangiectasia alone was not sufficient for a diagnosis of ETR. Participants defined as having TP had visible telangiectasia and mild or absent flushing as well as some prominent feature(s) of photodamage, such as an atrophic appearance, history of premalignant or malignant neoplasms, facial wrinkling, dyspigmentation, and/or poikiloderma; a key distinguishing factor was the absence of substantial or symptomatic flushing. Twenty-six participants had ETR; 20 had TP. Because the typical patient with ETR is a younger to middle-aged woman, while the typical patient with TP is an older man, participants were not age- or sex-matched. An additional 11 control participants, age- and sex-matched to ETR participants, were recruited. These controls lacked visible telangiectasia, redness, or other clinical signs of rosacea or TP.
Participants underwent a comprehensive medical history and physical examination. Primary and secondary features of rosacea, as outlined by the National Rosacea Society,4 were graded as absent, mild, moderate, or severe. Participants described extent, duration, and quality of flushing as well as flushing triggers. Physicians graded flushing, nontransient erythema, and telangiectasia on a scale of 0 to 3, with 3 being the most severe.4 Photodamage was graded using the Hamilton-Griffiths photonumeric scale (0 indicating no photodamage; 8, severe photodamage).5 Fitzpatrick skin phototype was determined,6 and facial erythema and pigmentation were recorded using a Chroma Meter CR-400 Colorimeter (Konica Minolta Sensing Inc). Five dermatology residents unfamiliar with the study’s diagnostic criteria evaluated participant photographs in a blinded fashion; they determined whether erythema was primarily central or lateral and whether these findings involved the nose. An interrater reliability analysis was performed to determine consistency among raters using SPSS Statistics software, version 22 (IBM Corporation).
Two 3-mm punch biopsy specimens were taken from the cheek of each participant. In ETR and TP participants, biopsies were centered on a visible telangiectatic vessel (Figure 1). In ETR, specimens were harvested whether or not the condition was in a flared state. One 3-mm biopsy specimen was immediately frozen in optimal cutting temperature (OCT) medium and used for immunohistologic and gene expression analysis. The second was used for light and electron microscopy.
One-half of a 3-mm biopsy specimen was fixed overnight in 4% buffered formalin, embedded in paraffin, sectioned, and stained with hematoxylin-eosin. Light microscopy sections were examined by a board-certified pathologist and imaged using a Zeiss Axio Imager, M1 microscope with the Axio Vision Imaging System (Zeiss). Each section was evaluated semiquantitatively or quantitatively with regard to several features, as listed in eTable 1 in the Supplement.
The other half of the split biopsy specimen was prepared and sectioned for transmission electron microscopy, as has been described previously.7,8
The OCT-embedded frozen tissue and formalin-fixed paraffin-embedded tissue were cut into 7-μm-thick sections and stained with monoclonal mouse antibody to human mast cell tryptase (M7052; Dako) by the immunoperoxidase method. To assess specificity of staining, substitution of isotype IgG1 for the primary antibody was used; no staining was visualized with the isotype. Image analysis was used to quantify the area of positive staining using commercially available software (Image-Pro; Media Cybernetics Inc). Immunostaining was quantified as a percentage of positive-staining area in the dermis per total dermal area. In parallel, stained cells were counted by an experienced pathologist to quantitate mast cell numbers in tissue sections (4 high-power fields per section).
After mRNA was extracted from OCT-embedded samples, specific gene transcripts were quantified by real-time reverse transcriptase polymerase chain reaction (PCR), as described elsewhere.9,10 Custom PCR primers and probes are listed in eTable 2 in the Supplement. All other primer-probe sets were validated gene expression assays (TaqMan; Applied Biosystems). Gene expression data were normalized to internal control housekeeping gene 36B4 transcript levels, and are presented as number-fold change in ETR skin specimens compared with control or TP skin specimens.
Categorical variables were summarized by frequency (number) and percentage for each response category. Continuous variables were summarized using mean, standard error (SE), and range. General linear models compared groups using 1-way analysis of variance with Benjamini-Hochberg multiplicity correction. Geometric mean number-fold changes in gene expression are presented following back-transformation of mean log number-fold changes on laboratory data obtained from statistical analysis. Statistical significance was determined using Kruskal-Wallis tests for ordinal data. The χ2 or Fisher exact tests were used to assess group differences for categorical data. Wilcoxon-Mann-Whitney post hoc tests, with Hochberg multiple comparison adjustments, were run on each significant pair. Pearson, Spearman, and point-biserial correlation coefficients were generated to determine the strength of parameter relationships.
The data were analyzed using SAS software, version 9.2 (SAS Institute Inc) with a 2-sided type I error rate set at .05.
The clinical and demographic features of the participants and all data supporting the following reported results, comparisons, and P values are included in the Table. Participants with TP demonstrated significantly more photodamage than those with ETR. This was expected because the TP population was slightly older, and photodamage was a diagnostic criterion for TP. Features characteristic of photodamage included an atrophic appearance, facial wrinkling, dyspigmentation, and poikiloderma of Civatte; all findings were significantly different from those found in the control groups except facial wrinkling. In addition, 70% of participants with TP (n = 14) had a history of skin cancer compared with 35% of those in the ETR group (n = 9) and 36% of controls (n = 4). Admittedly, there may have been some selection bias in terms of skin cancer frequency in the TP population because participants with a history of skin cancer were more willing to undergo facial biopsies. Population studies have found that rosacea is more common in women and often develops in the 30s or 40s. Our ETR population was older and predominantly male; our TP population was of a similar age and was entirely male. During participant recruitment, women and younger people were less willing to undergo facial biopsy.
The primary features of rosacea found in our study population and all data supporting the following reported results are included in the Table. Flushing was significantly greater in the ETR group than in the TP and control groups. This was expected because participants with moderate or severe flushing were considered to have ETR and were excluded from the TP and control groups. The mean (SD) Chroma Meter erythema scores in both the ETR and TP groups were significantly higher than among controls; erythema scores in the ETR and TP groups did not differ significantly from one another.
On analysis of multiple participant photographs by multiple blinded raters (full data reported in the Table), there was a strong relationship between location of erythema and diagnosis (79% of participants with ETR had primarily central erythema; 57% of participants with TP had primarily lateral involvement) (P < .001) (Figure 1). The interrater reliability coefficient was 0.752 (95% CI, 0.605-0.852) (P < .001). There was also a relationship between nasal involvement and diagnosis (80% with ETR had nasal involvement; 64% with TP had nasal involvement) (P = .01). The interrater reliability coefficient was 0.639 (95% CI, 0.436-0.782) (P < .001).
Figure 2 illustrates the histologic features of the 3 groups, and the Table includes all specific data points supporting the following reported results. Quiz Ref IDSignificantly more inflammation was observed in ETR than in TP (P = .02) and control groups (P = .001). Inflammation in both ETR and TP skin was widespread, whereas in control skin, if present, inflammation tended to be perifollicular. Vessels were more dilated in ETR specimens than in control specimens (P = .04); however, there was no statistical difference between TP and control (P = .07) or between ETR and TP (P = .74). In keeping with our clinical data showing greater photodamage in TP, there was significantly more collagen damage and solar elastosis in TP than in ETR (P = .01). Sebaceous glands were larger and more numerous in ETR than in control specimens (P = .01); in TP, there was a trend toward larger and more numerous sebaceous glands than in control samples, but this difference was nonsignificant (P = .06). Presence or absence of Demodex mites, spongiosis, and hyperkeratosis or parakeratosis did not differ significantly among the 3 groups.
Figure 3A is a composite image showing typical histologic features of ETR and TP specimens. Representative specimens from ETR and TP participants were assessed by transmission electron microscopy. Considerable connective tissue alterations were observed surrounding dilated vessels in TP. Collagen was thin and wispy, with a large amount of open space and clumps of amorphous material consistent with solar elastosis. Dilated vessels in ETR participants did not exhibit the same degree of surrounding collagen damage and generally were surrounded by thicker, intact collagen fibrils (Figure 3A).
Increased mast cell numbers have been noted in rosacea.11,12 On immunohistologic examination, mast cell counts were not significantly increased in either ETR or TP relative to control specimens (Table). However, quantitative image analysis of tryptase staining area demonstrated greater than 4-fold increased staining in ETR vs TP (P = .01) and 25-fold greater staining in ETR vs control (P < .001) (Figure 3C). Tryptase is stored within intracellular granules and secreted at mast cell degranulation; staining indicates presence of both intact and degranulated mast cells in tissues.13 The greater staining area observed in ETR therefore suggests increased degranulation rather than an increase in mast cell number (Figure 3B).
Our laboratory data, reported in full in the Table, indicate significantly greater flushing, inflammation, and mast cell activation and less photodamage in ETR than in TP. Next, we sought molecular correlates of these differences by examining gene expression of selected mast cell–activating neuropeptides, immune modulators, and extracellular matrix (ECM) components. We also examined expression of antimicrobial peptide genes, which have been implicated in the pathogenesis of rosacea.14 Fifteen genes were overexpressed in ETR samples compared with controls (eTable 3 in the Supplement). Of these genes, 10 were overexpressed in ETR compared with TP.
Compared with control samples, the level of calcitonin gene-related peptide (CGRP)-α was elevated 10-fold in ETR; CGRP-β was elevated 7-fold; and substance P was elevated 28-fold. Levels of CGRP-α and substance P were also significantly elevated in ETR compared with TP samples (9-fold and 5-fold, respectively) (Figure 4 and eTable 3 in the Supplement).
As detailed in the Table and eTable 3 in the Supplement, several cytokines and chemokines involved in inflammation and mast cell activation were assessed. Levels of tumor necrosis factor (TNF) and the chemokine CXCL12 and its receptor CXCR4 were significantly elevated in ETR compared with control samples: TNF was elevated 2-fold; CXCL12 was elevated 2.5-fold; and CXCR4 was increased 8-fold. Levels of TNF and CXCR4 were also significantly increased in ETR compared with TP (3-fold and 2.5-fold, respectively). Tumor necrosis factor is a potent activator of innate immunity, and CXCL12 is a multifunctional chemokine that acts as a chemoattractant and activator of mast cells through its receptor, CXCR4.
The most abundant structural protein in dermis is type I collagen, which forms fibrils that interact with type III collagen, proteoglycans, and matricellular proteins. Type I collagen makes up the bulk of the dermal ECM and provides structural support. As detailed in eTable 3 in the Supplement, which contains all data on which the following results are based, types I and III collagen and decorin (the most abundant proteoglycan in human dermis) were significantly overexpressed in ETR compared with control samples: types I (P < .001) and III (P = .001) collagen were elevated approximately 5-fold and decorin (P < .001), 10-fold. Relative to TP, type I collagen, type III collagen, and decorin were also significantly elevated (2-fold [P = .03], 3-fold [P = .02], and 8-fold [P < .001], respectively). Quiz Ref IDTurnover of collagen fibrils in human dermis is mediated by matrix metalloproteinases (MMPs)-1, -3, and -9, which are induced by injury and inflammation. Levels of all 3 MMPs were significantly elevated in ETR compared with control samples: MMP-1 (P = .01) and MMP-3 (P = .004) were elevated approximately 5-fold, and MMP-9 (P < .001) was elevated 33-fold. Also, MMP-3 was significantly elevated in ETR compared with TP (4-fold; P = .004).
Finally, we examined gene expression of several anti-microbial peptides (AMPs) and proteases involved in AMP processing. All data supporting the following reported results are in eTable 3 in the Supplement. Levels of defensin A1 was significantly elevated in ETR compared with control (P = .001) and TP (P = .001) samples (92-fold for each). However, levels of other AMPs, including HBD-2, HBD-3, and cathelicidin, as well as the processing enzymes stratum corneum chymotryptic enzyme (SCCE) and stratum corneum tryptic enzyme (SCTE), were not elevated in ETR compared with control tissue.
All data supporting the following reported results are in the Table. In ETR, more frequent flushing correlated with increased telangiectasia (r = 0.61; P = .003) and correlated strongly with the physician grades of telangiectasia (r = 0.96; P < .001) and with participant assessment of flushing severity (r = 0.67; P < .001). Participant assessment of flushing severity also correlated with physician assessments of nontransient erythema severity (r = 0.53; P = .01) and telangiectasia severity (r = 0.60; P = .003); it also correlated with physician grades of flushing (r = 0.83; P < .001), nontransient erythema (r = 0.55; P = .01), and telangiectasia (r = 0.73; P < .001). Finally, severity of ETR, as assessed by the physician, correlated well with physician grades of flushing (r = 0.61; P < .001), nontransient erythema (r = 0.66; P < .001), and telangiectasia (r = 0.73; P < .001). A burning or stinging sensation correlated with mRNA levels of substance P (r = 0.65; P = .04).
Facial telangiectasia develops in both rosacea and photodamage1,15,16; however, mechanisms underlying vessel dilation in both conditions are poorly understood. Previous research has demonstrated that an MMP-degraded collagen fibril framework promotes formation of tubelike vascular structures by endothelial cells.17 Extensively fragmented collagen fibrils associated with photodamage may permit development of the dilated, telangiectatic vessels seen in TP. It should be noted in this regard that we observed upregulation of MMP gene expression in both TP and ETR participants compared with controls. In fact, expression of several matrix remodeling genes was more elevated in ETR than in TP samples, despite greater clinical and histologic evidence of connective tissue alteration in TP. Upregulation of matrix remodeling gene expression has been described previously in rosacea, most notably in phymatous and papulopustular rosacea but also in ETR.12 It is possible that ECM remodeling, driven by the inflammatory milieu characteristic of ETR, directly affects vascular cell biology or produces subtle changes in connective tissue that permit vessel dilation.
A distinct clinical feature of ETR is flushing. It has been proposed that repeated episodes of vasodilation associated with flushing lead to loss of vascular tone and permanent vessel dilation. These dilated vessels may become leaky, with subsequent release of inflammatory mediators and continued cutaneous inflammation and telangiectasia.18,19 Within our ETR population, a history of more frequent flushing correlated with increased telangiectasia. This finding is consistent with the theory that repeated flushing episodes eventually lead to permanent telangiectasia.
The flushing of rosacea is often uncomfortable. Patients complain of a hot feeling and sometimes have associated burning or stinging. Underlying events contributing to flushing in rosacea are not fully understood. In our gene expression analysis, several neuropeptide genes that could contribute to flushing were upregulated in ETR: CALCA (CGRP-α), CALCB (CGRP-β), and TAC1 (substance P). Substance P acts on vascular smooth muscle to increase production of nitric oxide, with subsequent vasodilatation and increased vascular permeability.20 A role for substance P in rosacea has long been suggested.21- 23
One of the most potent microvascular vasodilators yet identified is CGRP (over 100 times more potent than substance P).24 Both CGRP isoforms, -α and -β, have similar activity, though CGRP-α is the predominant isoform in human skin and colocalizes with substance P.24 Recent evidence suggests that CGRP-β is produced by epidermal keratinocytes and upregulated in certain cutaneous pain syndromes.25Quiz Ref ID Upregulation of substance P and CGRP may be mediated through activation of nonselective cation channels TRPV1 and TRPA1 (transient receptor potentials vanilloid 1 and ankyrin 1); agonists of TRPV1 and TRPA1 have been shown in mice to cause increased blood flow via upregulation of vasodilator neuropeptides such as substance P and CGRP.26 Certain stimuli (eg, heat, cold, spices) may lead to activation of TRPV1 and TRPA1, with subsequent upregulation of substance P and CGRP and further downstream vasodilation, producing the typical flush of rosacea.26,27
Quiz Ref IDNeuropeptides may also affect mast cell function. Substance P can bind to mast cells and cause degranulation,28 with release of histamine and 5-hydroxytryptamine, which bind to vascular receptors and cause vasodilation and increased vascular permeability.29 Increased mast cell numbers have been demonstrated in rosacea by several groups.11,12 Although we demonstrated no statistically significant differences in mast cell numbers between ETR and TP participants, we did observe a trend toward higher numbers in ETR vs TP and control samples. More importantly, we observed increased mast cell degranulation in ETR as compared to TP and control. The chemokine CXCL12 and its receptor CXCR4, which play a role in mast cell migration,30 were also more highly expressed in ETR. Although the complicated interplay of various cellular and molecular pathways is still not completely clear, our data suggest important roles for mast cells and neuropeptides in ETR.
In recent years, emphasis has been placed on the role of antimicrobial peptides in rosacea, specifically upregulation of cathelicidin and increased levels of SCCE and SCTE, which cleave cathelicidin to form a small peptide, LL-37.14 Our gene expression data show no significant difference in cathelicidin, SCCE, or SCTE mRNA levels among participants with ETR or TP vs control. It should be noted that Yamasaki et al14 examined samples from the nasomalar fold of patients with papulopustular rosacea (R.L. Gallo, MD, PhD; written communication; October 2013), whereas our samples were taken from the cheeks of participants with ETR. However, another study demonstrated increased gene expression of cathelicidin not only in papulopustular rosacea but also in phymatous rosacea and ETR12; the facial skin examined there was from previously obtained diagnostic biopsies, but skin site was not mentioned.12 Recent data indicate that different sites on the skin, even on the face, can vary considerably in the bacterial diversity of the overlying microbiome31; highly sebaceous sites have less diversity, while drier sites have more diversity.32 Bacteria native to a site play a key role in educating antimicrobial peptides in that site.33,34 Therefore, the site chosen for a biopsy can influence the findings of not only local bacteria but also antimicrobial peptides. When comparing various studies examining antimicrobial peptide expression in rosacea, it is imperative that the biopsy site be noted. In addition, the field would benefit considerably from a study examining variability of antimicrobial peptide production in normal skin.
Because the typical patient populations of ETR and TP differ, groups in the present study were not age- or sex-matched. Despite this, both participant populations were entirely or primarily male, and average participant ages were similar. Certain criteria used for diagnosis, such as flushing and photodamage, necessarily resulted in increased flushing in ETR and increased photodamage in TP. Finally, opinion varies considerably, even among rosacea experts, as to distinguishing ETR from telangiectatic changes typical of photodamage; we made our best attempt to separate out these groups according to various criteria, but there is no established consensus. In all likelihood, sometimes overlap occurs. Given these difficulties and potential shortcomings, it is even more striking that significant molecular differences emerged between the groups.
We have described herein a form of photodamage characterized by significant erythema and telangiectasia, which we term telangiectatic photoaging (TP). This condition is distinct from ETR—clinically, histologically, and in gene expression pattern. Vessels are dilated in both ETR and TP. In TP, the visibly altered collagen fibrils surrounding the vessels may permit vessel dilation, whereas in ETR, subhistologic alteration in the ECM, with upregulation of MMPs, may contribute to vessel dilation. The data presented herein also support a role for mast cells in ETR and suggest key roles for the neuropeptides substance P and CGRP-α.
Corresponding Author: Yolanda R. Helfrich, MD, Department of Dermatology, University of Michigan Medical School, 1500 E Medical Center Dr, 1910 Taubman Center, SPC 5314, Ann Arbor, MI 48109-5314 (firstname.lastname@example.org).
Published Online: March 23, 2015. doi:10.1001/jamadermatol.2014.4728.
Author Contributions: Dr Helfrich had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Helfrich, Fisher, Sachs, Varani, Voorhees.
Acquisition, analysis, or interpretation of data: Helfrich, Maier, Cui, Fisher, Chubb, Fligiel, Voorhees.
Drafting of the manuscript: Helfrich, Cui.
Critical revision of the manuscript for important intellectual content: Helfrich, Maier, Fisher, Chubb, Fligiel, Sachs, Varani, Voorhees.
Statistical analysis: Chubb.
Obtained funding: Helfrich.
Administrative, technical, or material support: Helfrich, Fisher.
Study supervision: Helfrich.
Conflict of Interest Disclosures: Dr Helfrich served as a consultant to Galderma. Dr Maier was the principal investigator for work supported by a grant from Valeant Pharmaceuticals and receives royalties from UpToDate.
Funding/Support: This study was supported in part by the Dermatology Foundation (Dr Helfrich), National Rosacea Society (Dr Helfrich), and the University of Michigan Department of Dermatology.
Role of the Funder/Sponsor: The funding institutions had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Previous Presentation: This research was presented at the Annual Meeting of the American Academy of Dermatology; March 23, 2015; San Francisco, California.
Additional Contributions: Craig Hammerberg, PhD, assisted with immunostaining. Kathryn Keeley, BS, Jennifer Bell, BA, and Harrold Carter, BS, assisted with data collection and photography. Laura Vangoor, BFA, assisted with figures and tables. Jackie Giletto, BA, helped edit for grammar and language flow. No contributor received compensation for their work on this article beyond that provided in the normal course of their employment.