palmoplantar pustulosis: a clinical and immunohistological study

9
Palmoplantar pustulosis: a clinical and immunohistological study M-O.ERIKSSON, E.HAGFORSEN, I.P.LUNDIN AND G.MICHAE ¨ LSSON Department of Dermatology and Venereology, University Hospital, S-751 85 Uppsala, Sweden Accepted for publication 17 October 1997 Summary Pustulosis palmoplantaris (PPP) is a common chronic skin disease, which is very resistant to treatment. It is not known why the lesions are located in the palms and soles. There are few studies of the disease and in particular studies of the histology. Fifty-nine patients with PPP answered a questionnaire concerning their medical history and 39 of them were clinically examined. Biopsy specimens were taken from involved skin in 22 of the 39 patients and studied immunohistologically for tryptaseþ mast cells, EG2þ eosinophils, lipocalinþ neutrophils and CD3þ T lymphocytes. The sweat gland and sweat duct were visualized with AE1/AE3 antibody (cytokeratins 1–8, 10, 14/15, 16, 19). In addition to neutrophils in the pustule and lymphocytes in the upper dermis, there were also large numbers of mast cells and eosinophils in the subpustular area. Numerous eosinophils were present in the pustule. The epidermal part of the eccrine duct was not detectable in any of the specimens from patients with PPP but was present in all of the nine control persons (including two smokers). The results indicate that the acrosyringium is involved in the inflammation and also that mast cells and eosinophils participate in a hitherto unknown way. Of the 39 patients clinically examined, two had previously diagnosed thyroid disease and two had gluten hypersensitivity. Seventeen had one or several abnormal serum concentrations of thyroid-stimulating hormone, thyroxin, antibodies against thyroglobulin or thyroperoxidase and 10 had immunoglobulin (Ig) A antibodies to gliadin. The mean 6 SD for serum IgA and for eosinophil cationic protein was increased. From the questionnaire the most notable finding was that 56 of the 59 patients had been or still were smokers, all of whom had started smoking before the first signs of PPP. We hypothesize that the acrosyringium might be the target for the inflammation and that PPP is linked to autoimmune thyroid disease and smoking. Palmoplantar pustulosis (PPP) is a common chronic skin disease characterized by sterile intraepidermal pustules and usually also erythematous, scaly skin. PPP is more common in women than in men and is also more common among smokers than in non- smokers. 1–3 An association with thyroid disease has been reported. 1 Patients with PPP may occasionally have thin psoriasis-like lesions on the forearms and legs, but the relationship of PPP to psoriasis is contro- versial. However, we have previously found similar neutrophil dysfunctions in PPP and psoriasis, as well as similar chemokinetic effects of sera from patients with these two diseases. 4 In addition, in both conditions, the serum level of eosinophilic cationic protein was elevated. 5 Why the PPP lesions are located in the palms and soles is not known. In comparison with other skin regions palms and soles are very rich in eccrine sweat glands but their appearance in PPP has not been studied. In this paper we report on anamnestic, clinical and laboratory data in PPP patients, with particular emphasis on previously unknown immuno- histological features of the inflammation and the sweat gland apparatus. Patients and methods Questionnaire Fifty-nine patients (52 women, 21–79 years old, and seven men, 43–71 years old), with typical PPP of the hands and/or feet, answered a questionnaire concern- ing their medical history with particular emphasis on their skin disease. Questions about smoking, including age at the start of smoking and the approximate number of cigarettes/day during each decade were also asked. British Journal of Dermatology 1998; 138, 390–398. 390 q 1998 British Association of Dermatologists Correspondence: Professor Gerd Michae ¨lsson.

Upload: eriksson

Post on 06-Jul-2016

215 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Palmoplantar pustulosis: a clinical and immunohistological study

Palmoplantar pustulosis: a clinical and immunohistologicalstudy

M-O.ERIKSSON, E.HAGFORSEN, I.P.LUNDIN AND G.MICHAELSSONDepartment of Dermatology and Venereology, University Hospital, S-751 85 Uppsala, Sweden

Accepted for publication 17 October 1997

Summary Pustulosis palmoplantaris (PPP) is a common chronic skin disease, which is very resistant totreatment. It is not known why the lesions are located in the palms and soles. There are few studies ofthe disease and in particular studies of the histology. Fifty-nine patients with PPP answered aquestionnaire concerning their medical history and 39 of them were clinically examined. Biopsyspecimens were taken from involved skin in 22 of the 39 patients and studied immunohistologicallyfor tryptaseþ mast cells, EG2þ eosinophils, lipocalinþ neutrophils and CD3þ T lymphocytes. Thesweat gland and sweat duct were visualized with AE1/AE3 antibody (cytokeratins 1–8, 10, 14/15,16, 19). In addition to neutrophils in the pustule and lymphocytes in the upper dermis, there werealso large numbers of mast cells and eosinophils in the subpustular area. Numerous eosinophils werepresent in the pustule. The epidermal part of the eccrine duct was not detectable in any of thespecimens from patients with PPP but was present in all of the nine control persons (including twosmokers). The results indicate that the acrosyringium is involved in the inflammation and also thatmast cells and eosinophils participate in a hitherto unknown way. Of the 39 patients clinicallyexamined, two had previously diagnosed thyroid disease and two had gluten hypersensitivity.Seventeen had one or several abnormal serum concentrations of thyroid-stimulating hormone,thyroxin, antibodies against thyroglobulin or thyroperoxidase and 10 had immunoglobulin (Ig) Aantibodies to gliadin. The mean 6 SD for serum IgA and for eosinophil cationic protein wasincreased. From the questionnaire the most notable finding was that 56 of the 59 patients hadbeen or still were smokers, all of whom had started smoking before the first signs of PPP. Wehypothesize that the acrosyringium might be the target for the inflammation and that PPP is linkedto autoimmune thyroid disease and smoking.

Palmoplantar pustulosis (PPP) is a common chronicskin disease characterized by sterile intraepidermalpustules and usually also erythematous, scaly skin.PPP is more common in women than in men and isalso more common among smokers than in non-smokers.1–3 An association with thyroid disease hasbeen reported.1 Patients with PPP may occasionallyhave thin psoriasis-like lesions on the forearms andlegs, but the relationship of PPP to psoriasis is contro-versial. However, we have previously found similarneutrophil dysfunctions in PPP and psoriasis, as wellas similar chemokinetic effects of sera from patientswith these two diseases.4 In addition, in both conditions,the serum level of eosinophilic cationic protein waselevated.5 Why the PPP lesions are located in thepalms and soles is not known. In comparison withother skin regions palms and soles are very rich in

eccrine sweat glands but their appearance in PPP hasnot been studied. In this paper we report on anamnestic,clinical and laboratory data in PPP patients, withparticular emphasis on previously unknown immuno-histological features of the inflammation and the sweatgland apparatus.

Patients and methods

Questionnaire

Fifty-nine patients (52 women, 21–79 years old, andseven men, 43–71 years old), with typical PPP of thehands and/or feet, answered a questionnaire concern-ing their medical history with particular emphasis ontheir skin disease. Questions about smoking, includingage at the start of smoking and the approximate numberof cigarettes/day during each decade were also asked.

British Journal of Dermatology 1998; 138, 390–398.

390 q 1998 British Association of Dermatologists

Correspondence: Professor Gerd Michaelsson.

Page 2: Palmoplantar pustulosis: a clinical and immunohistological study

Clinical examination

Thirty-nine of the patients (35 women, four men) whoanswered the questionnaire were examined clinically(G.M.). The degree of erythema and scaling was gradedfrom 0 to 4. The number of fresh pustules on each handand foot was counted, and the involved areas weremarked on a template.

None of the patients were taking beta-blockers orlithium. One patient was on a low dosage of etretinate(20 mg daily) and also used betamethasone topically.Although most patients considered topical treatment oflimited value, 17 used topical corticosteroids occasion-ally, when new pustules appeared; eight used calcipo-triol during periods with scaly skin, and six used coal tar5% in zinc oxide paste. The majority of the patients wereonly using emollients at the time of the examination.

Blood samples

Triiodothyronine (T3), thyroxine (T4) and thyroid-stimulating hormone (TSH) in serum were measuredby routine methods employed in the department ofclinical chemistry (normal ranges T3 1·2–2·8 nmol/L,T4 9–21 nmol/L, TSH 0·3–4·0 mU/L). Serum anti-bodies to thyroid globulin and thyroid peroxidase weredetermined by haemagglutination (titres lower than1/10 and 1/640 were considered normal), and anti-bodies to parietal cells by indirect immunofluorescenceaccording to the clinical routine (titre lower than 1/25normal). Serum immunoglobulin (Ig) A and IgG anti-gliadin antibodies (IgA AGA and IgG AGA) were deter-mined with an ELISA method as previously described.6

IgA AGA >50 U/mL and IgG AGA >12 U/mL wereconsidered to be elevated. Serum IgA, IgG, IgM andIgE were measured by routine methods. Serum eosino-phil cationic protein (ECP) was measured with a radio-immunoassay; values >16 mg/L were consideredelevated.7 The results of the immunoglobulin and ECPdeterminations were compared with those in healthyblood donors.

Skin biopsies

After intradermal injection of xylocaine-adrenalinepunch biopsy specimens (3 mm in diameter) weretaken from involved skin either from the hypothenarregion, the thenar region or the lateral or medialaspects of the heel. In four patients biopsy specimenswere also taken from seemingly non-involved hypo-thenar skin. One to three specimens were taken fromeach patient and fixed in buffered formalin 4% or

snap-frozen. For comparison, biopsy specimens fromtwo women with pronounced keratotic tylotic eczemaof the palms were also included. Seven women 24–65years old and two men 24 and 68 years old, all healthy,served as a reference group; one of the women, aged55 years old, smoked 10 cigarettes per day and one man20 cigarettes per day, both had smoked for many years.The biopsy specimens from the reference people weretaken from the hypothenar region.

Immunohistological examination of frozen biopsyspecimens

Cryostat sections, 6 mm thick were stained by meansof a three-step monoclonal antibody peroxidase anti-peroxidase (PAP) technique. CD3þ T lymphocytes werestained with antihuman Leu-4 (dilution 1/100; Becton-Dickinson, Oxnard, CA, U.S.A.), and eosinophils withantihuman EG-2, which binds to ECP (dilution 1/200;Kabi Pharmacia, Uppsala, Sweden). The sections werefixed with acetone. Endogenous peroxidase activity wasblocked by incubation in 0·3% H2O2 in phosphate-buffered saline (PBS) for 15 min. The sections werethen allowed to react with normal rabbit serum (dilu-tion 1/10) for 10 min to reduce non-specific staining.Thereafter they were incubated for 60 min with theprimary antibody. Rabbit antimouse IgG (dilution 1/40;Dakopatts, Glostrup, Denmark) was used as a secondaryantibody. Finally, the sections were incubated in a thirdstep with horseradish peroxidase–mouse monoclonalantihorseradish peroxidase (dilution 1/250; Dakopatts).The peroxidase reaction was developed with 3-amino-9-ethylcarbazole.

To exclude the possibility of false positive stainingdue to endogenous peroxidase from the neutrophils,the alkaline phosphatase–antialkaline phosphatase(APAAP) technique was also used to visualize eosino-phils. In that case, the cryostat sections were incubatedwith normal rabbit serum, and with primary andsecondary antibodies as described above. In the thirdstep they were incubated with the mouse APAAPcomplex (dilution 1/25; Dakopatts). The phosphatasewas visualized with DAKO Fast Red Substrate system(Dakopatts).

Neutrophils were stained with a monoclonal anti-human neutrophil lipocalin (anti-HNL) antibody (dilu-tion 10 mg/mL, obtained as a kind gift from PharmaciaDiagnostics, Uppsala, Sweden).8 The sections were fixedin cold methanol for 60 min, incubated with normalrabbit serum and thereafter with primary antibodyfollowed by the APAAP complex as described above.

PALMOPLANTAR PUSTULOSIS 391

q 1998 British Association of Dermatologists, British Journal of Dermatology, 138, 390–398

Page 3: Palmoplantar pustulosis: a clinical and immunohistological study

The sections were counterstained with Mayer’s haemat-oxylin. Controls with mouse IgG (Dakopatts) in thesame dilution as the primary antibodies were made,and all were negative.

Immunohistological examination of paraffin-embeddedbiopsy specimens

Paraffin sections, 5 mm thick, were deparaffinized andrehydrated. After blockage of endogenous peroxidasewith 3% H2O2 in methanol for 15 min, the sections wereallowed to react with 10% normal rabbit serum for10 min. The sections were then incubated overnight at4 8C with the primary antibody monoclonal antibody(mAb) 1222 (dilution 1/5000; Chemicon InternationalInc., Temecula, CA, U.S.A.),9 which binds to mast celltryptase. Thereafter the same procedure as described forthe PAP technique for frozen biopsy specimens wasfollowed.

The sweat gland and its duct were stained with apooled mouse monoclonal antibody to human epithelialkeratins, antikeratin AE1 (cytokeratins 10, 14/15, 16,19)/AE3 (cytokeratins 1–8), (dilution 1/1000; Boeh-ringer Mannheim Corp., Indianapolis, IN, U.S.A.).10

After blockage of endogenous peroxidase, as describedabove, the sections were treated with 0·05% proteasefor 10 min and thereafter the same procedure wasfollowed as for mAb 1222. The sections were counter-stained with Mayer’s haematoxylin. Controls withmouse IgG (Dakopatts) in the same dilution as theprimary antibodies were made, and all were negative.

Evaluation of the histopathological findings

All evaluations were made on coded sections. Thenumber of cells was counted in the epidermis, theupper dermis below the pustule, the upper dermis withexclusion of the area below the pustule, and the lowerdermis as shown in Fig. 1. To define the upper and lowerdermis, a grid was used in which one unit measured0·21 × 0·21 mm. The upper dermis was defined as oneunit below the epidermis (Fig. 1). Usually the cells werecounted in at least three sections and the mean numberper section was calculated. The number of cells in thepustule was evaluated semiquantitatively. The sweatgland and the duct were studied with special emphasison the appearance of the epidermal part of the duct.

Statistics

StatView II (Abacus Concepts, Berkeley, CA, U.S.A.) was

used for paired (Wilcoxon signed rank test) or non-paired (Mann–Whitney U-test) non-parametric tests forcomparison of data and for all other calculations.

Results

The questionnaire

Most commonly the first PPP symptoms appearedbetween the ages of 30 and 50 years (mean 6 SD42 6 14 years), but there was a range from 15 to 66years. The duration of the disease varied between 1 and56 years. Fifty of the 59 patients had only been free fromtheir PPP for short periods and never for more than1 year, whereas nine had had one or several periods of1–6 years without symptoms. Some of the answers tothe questionnaires are summarized in Table 1. Thefigures reported for heredity for psoriasis, PPP, thyroiddisease and gluten intolerance include cases amongparents, siblings and children. In addition, some otherrelatives were known to be affected (see bottom ofTable 1).

Six of the patients thought that they had bothpsoriasis and PPP, but they considered their psoriasisas mild, with only a few lesions. Arthralgia was reportedby 25 patients, in 18 of whom it was located in thehands and fingers. Fifty-six of the 59 patients reportedthat the skin lesions were itching, particularly whennew pustules were under development. At the onset ofthe PPP 56 of 59 patients were smokers. Sixteen hadstopped smoking or had reduced the number of cigar-ettes in recent years, but none of the patients wereaware of any influence of their smoking habits on the

392 M-O.ERIKSSON et al.

q 1998 British Association of Dermatologists, British Journal of Dermatology, 138, 390–398

Figure 1. Schematic illustration of the different sections used toevaluate the number of cells in the specimens and referred to inTable 3. A ¼ epidermis; B ¼ upper dermis, below pustule; C ¼ upperdermis; surrounding the subpustular area; D ¼ lower dermis.

Page 4: Palmoplantar pustulosis: a clinical and immunohistological study

PPP symptoms. As shown in Table 1, in addition to theirPPP many patients had a history of other disorders,in particular thyroid disease, gluten intolerance anddiabetes. A history of allergic rhinitis was reported by10 patients.

Clinical examination

Three of the 39 patients who underwent clinical exami-nation had PPP only on the hands (one or both), andeight only on the feet. In 27 patients both the hands andfeet were involved. One patient had been healed for1 month. Fresh pustules on the hands/feet wereobserved in 26 patients; the number of pustules onthe hands and feet varied widely (0–100). Most patientswith pustules had between three and 10 pustules on thehands and between five and 20 on the feet. The numberof pustules on the right or left foot was significantlyhigher than that on the right and left hand (P<0·04 inboth cases). Erythema and scaling were present in allbut one patient, and tended to be more pronounced onthe feet than on the hands (data not shown). Two or

three discrete, thin psoriasis-like lesions 1–2 cm indiameter were observed on the elbows and distal partsof the legs in a few patients. The numbers (mean 6 SD)of all pustules in patients smoking 0–5, 6–10 and >10cigarettes/day, were 9 6 14 (n ¼ 7), 16 6 15 (n ¼ 16)and 40 6 98 (n ¼ 16), respectively, at the time of exami-nation, but the differences were not statistically signifi-cant. A few patients had stopped smoking or cut downthe number of cigarettes in the last few years butthe number of patients was too small to permit anystatistical comparison with those with unchangedsmoking habits.

Blood samples

Seventeen of the 39 patients examined clinically hadone or several abnormal results concerning TSH, T4 orthyroid antibodies (TSH > 4·0 mU/L in four patients,TSH < 0·3 mU/L in two patients; T4 > 21 pmol/L in onepatient, < 9 pmol/L in six patients; thyroglobulin anti-body titre 1/10 in three patients, 1/80 in two patients;thyroperoxidase antibody titre 1/100 in three patients,1/400 in one patient, 1/1600 in three patients, 1/4000in one patient, 1/6400 in one patient, 1/124,000 in onepatient). Two of these 17 patients had thyroid disease,hypothyroidism and hyperthyroidism, respectively, thathad been diagnosed previously.

Two patients had parietal cell antibody titre >1/25.Ten patients had IgA antibodies to gliadin >50 U/L(50–70 kU/L in three patients, >70 U/L in sevenpatients), two of whom were known to have glutenhypersensitivity (subtotal and partial villous atrophy,respectively). Four of the 10 patients with IgA anti-bodies to gliadin also had thyroid antibodies and/orabnormal TSH or T4.

Women with PPP had a significantly elevated mean6 SD serum IgA value (P<0·0005) compared with thatin female blood donors: 2·70 6 0·9 g/L (n ¼ 34) vs.2·10 6 0·80 g/L (n ¼ 128). The mean serum IgMvalue was significantly lower in women with PPP(1·2 6 0·70 g/L, n ¼ 33) than in female blood donors(1·60 6 0·70 g/L, n ¼ 119), P<0·0025. The meanvalue for serum IgG was not significantly elevated inPPP: 10·8 6 2·9 g/L compared with 10·4 6 2·3 g/L infemale blood donors.

The mean 6 SD value for ECP in serum (both femaleand male patients) was 15·8 6 10·8 mg/L (range 4·4–56, n ¼ 36), compared with 6·8 6 3·6 mg/L (n ¼ 101) inblood donors (P<0·0005). Patients without a history ofallergic rhinitis (n ¼ 28) had an ECP value of 17·2 6

11·8 mg/L and those with such a history (n ¼ 8) had a

PALMOPLANTAR PUSTULOSIS 393

q 1998 British Association of Dermatologists, British Journal of Dermatology, 138, 390–398

Table 1. Anamnestic data in 59 patients (52 women, seven men) withpalmoplantar pustulosis (PPP) and inheritance among parents,siblings and children

Number (total n ¼ 59)

Heredity for:palmoplantar pustulosis 8psoriasis 13thyroid disease 13a

gluten intolerance 2b

Patients with history of:psoriasis 6thyroid disease 8gluten intolerance 5diabetes 4vitiligo 3alopecia areata 2

Precipitating factors first episode of PPPtrauma 7stress 15

Worsening associated with:some food items 5upper respiratory infections 12hot weather 21stress 27

Smoker at onset of PPP 56Pruritus 56Arthralgia 25History of sick leave due to PPP 26Difficulties in working because of PPP 41

a In addition 3 and b 2 nephews, aunts or uncles were affected.

Page 5: Palmoplantar pustulosis: a clinical and immunohistological study

value of 10·7 6 2·9 mg/L. The mean 6 SD for serum IgEwas 89 6 143 kU/L in those without a history of allergicrhinitis and 131 6 150 kU/L in those with.

Immunohistology

Mast cells. The number of mast cells was increased inthe upper dermis in involved PPP skin (n ¼ 19) incomparison with that in control subjects (n ¼ 9),(P ¼ 0·05). In specimens with a pustule the mast cellswere located below the pustule (Table 2, Fig. 2a),whereas they were less increased in the dermis sur-rounding the subpustular area. However, large numbersof mast cells were also present in the upper dermis insome of the specimens from involved skin withoutpustules. The number of mast cells was also significantlyincreased in the lower dermis compared with thenumber in the control group (P ¼ 0·04).

Eosinophil granulocytes. There were large numbers of

eosinophils in specimens containing a pustule. Thecontents of some pustules had been lost during thepreparation of the specimens, but in all of the pustulesstill containing cells, numerous eosinophils were pre-sent and many eosinophils were observed in the upperdermis below the pustule (Table 3, Fig. 2b). The speci-mens from non-involved skin and those from the controlgroup contained no or only few eosinophils in thedermis.

Neutrophil granulocytes. The pustules were located in theupper part of the epidermis. In those with a preservedcell content, there was a strong reaction to lipocalin,indicating the presence of neutrophils (Fig. 2c). Neutro-phils were also found in the upper dermis (mean 6 SD;1175 6 1672/mm2 in the subpustular area) and epi-dermis adjacent to the pustule. In the control group, noor very few neutrophils were observed.

CD3þ T lymphocytes. Large numbers of CD3þ T

394 M-O.ERIKSSON et al.

q 1998 British Association of Dermatologists, British Journal of Dermatology, 138, 390–398

Table 2. Numbers of tryptase þ mast cells, EG2þ eosinophils and CD3þ T lymphocytes/mm2 in patients with pustulosispalmoplantaris (PPP) and in healthy subjects

n A B C D

Mast cellsInvolved pustular skin 8 0 132 6 102 23 6 36 39 6 11

(36–318) (0–108) (27–61)Involved non-pustular skin 11 0 – 113 6 75 43 6 39

(23–266) (11–120)Non-involved PPP skin 4 0 – 9 6 18 0

(0–33)Healthy subjects, smokers 2 0 – 34 6 9 5 6 5

(27 ¹ 39) (2 ¹ 8)Healthy subjects, non-smokers 7 0 – 72 6 32 25 6 16

(41–127) (9–54)Healthy subjects, all 9 0 – 64 6 36 20 6 16

(27–127) (3–54)Eosinophil granulocytes

Involved pustular skin 11 0 48 6 82 9 6 14 9 6 29(0–81) (0–34) (0–88)

Involved non-pustular skin 12 0 – 5 6 9 0 6 2Non-involved PPP skin 4 0 – 0 0Healthy subjects, all 9 0 – 0 0

LymphocytesInvolved pustular skin 8 75 6 111 1213 6 753 413 6 236 57 6 59

(7–260) (517–2260) (0–655) (18–133)Involved non-pustular skin 11 34 6 32 456 6 377 38 6 25

(5 6 106) (133–1537) (5–75)Non-involved PPP skin 4 5 6 3 122 6 75 11 6 11

(0–7) (50–2260) (2–23)Healthy subjects, all 9 5 6 11 – 67 6 51 24 6 33

(0–23) (23–158) (0–45)

Mean 6 SD (range); n ¼ number. A ¼ epidermis, B ¼ upper dermis, below pustule, C ¼ upper dermis surrounding B and/or upperdermis in specimens without pustules, D ¼ lower dermis (Fig. 1).

Page 6: Palmoplantar pustulosis: a clinical and immunohistological study

PALMOPLANTAR PUSTULOSIS 395

q 1998 British Association of Dermatologists, British Journal of Dermatology, 138, 390–398

Figure 2. (a–f) Findings in skin biopsy specimens from pustular skin in patients with palmoplantar pustulosis (PPP). (a) Pronounced increase oftryptase þ mast cells in the papillary dermis below a pustule in stratum corneum and strong staining of tryptase in the papillary dermis, which is notpresent in healthy controls. (b) Pustule filled with EG2þ eosinophils. Also note eosinophil infiltration of the papillary dermis. In healthy controls noeosinophils were observed. (c) Lipocalin þ neutrophils in pustule and upper dermis. No neutrophils were observed in the controls. (d) CD3þ Tlymphocytes, arranged in groups in the upper dermis below the pustule. Lymphocytes are also present in the epidermis. (e) Staining of the sweatgland apparatus with the AE1/AE3 antibody in healthy controls. The dermal part of the duct stains dark brown. The pattern in stratum corneum ischaracteristic with regular red-brown columns. (f) Absence of the normal columnar pattern of the corneal part of the sweat gland duct in PPP. Notealso the strong staining of the epidermis. (a–f Original magnification × 10.)

Page 7: Palmoplantar pustulosis: a clinical and immunohistological study

lymphocytes were observed, often in groups, mainly inthe upper dermis, increased (P ¼ 0·001) compared withthe number in the controls (Table 2, Fig. 2d). This cellinfiltration was particularly pronounced below thepustule. Lymphocytes were present in the epidermis inall specimens in the PPP group, whereas this was onlyobserved in one of the controls. No CD3þ T lympho-cytes were seen in the pustules.

The sweat-gland apparatus. In healthy subjects, theeccrine ducts in the stratum corneum formed regularspiral columns (Fig. 2e). This columnar pattern was alsonoted in the two healthy control subjects who smoked.In specimens from healthy subjects discrete stainingof the duct was also observed in the basal layer ofthe epidermis. The ‘columnar’ pattern in the stratumcorneum was not seen in any of the specimens frominvolved PPP skin (Fig. 2f). In the specimens from thetwo patients with keratotic tylotic eczema of the hand,the intracorneal sweat ducts were normal.

The sweat glands and the dermal part of the sweatduct seemed to be of the same appearance in the PPPpatients as in the controls. Except a weak staining of thebasal cell layer, normal epidermis did not stain withAE1/AE3. In contrast, the PPP epidermis in involvedskin, always acanthotic, showed diffuse, strong staining.As the acrosyringium could not be detected, it was notpossible to determine the relation between the pustuleand the duct.

Discussion

There have been few immunohistological studies onPPP. The dominating feature in the histopathology ofPPP is the pustule which is filled with neutrophils.11

There is also a dermal lymphocyte infiltration. In thepresent study we have found several previouslyunknown features. For instance there were numeroustryptaseþ mast cells in the upper dermis in PPP skin. Inspecimens containing pustules it was evident that mastcells had accumulated below the pustule, whereas thenumber in areas surrounding the subpustular area wasonly slightly increased or not at all. Another newfinding was the presence of numerous EG2þ eosinophilsin the upper dermis below the pustules, whereas only afew eosinophils were observed in areas around the sub-pustular area. Furthermore, the pustules were foundto contain large numbers of eosinophils as well aslipocalinþ neutrophils, indicating that eosinophilsparticipate in the pustule formation together withneutrophils.

The dermal distributions of mast cells and eosinophilswere similar, but only eosinophils and neutrophils werefound within the pustules, which were located in theupper part of the epidermis. The reasons for the pro-nounced accumulation of mast cells and of neutrophilsand eosinophils are not known, but one importantchemoattractant for both neutrophils12 and eosino-phils13 is interleukin (IL) 8, which has been shown tobe present in the eccrine duct in general and in PPPepidermis.14 Furthermore, mast cells have been shownto secrete IL-8.15 The elevated serum levels of eosinophilcationic protein that were observed in many patientsprobably reflect an activation of eosinophils in the skin.We have previously found raised serum ECP and anincreased number of EG2þ eosinophils in active psori-asis,5 but the number of eosinophils in pustular areas inPPP seems to be larger than in psoriasis. There are alsonumerous mast cells in lesional skin in psoriasis.16 Thusthere are several similarities between psoriasis and PPPwith regard to the cellular components in the inflam-mation. However, psoriasis is usually not located in thepalms or soles. Lymphocytes had a more diffuse distri-bution than eosinophils and mast cells but with atendency to subpustular accumulation. The accumula-tion of the participating cells below the pustules, but notin the areas around, may indicate that there is anepidermal target for the inflammation which is notevenly distributed in the epidermis.

No studies have been focused on the appearance ofthe sweat gland and its duct in PPP. The antibody AE1/AE3 has been reported to be the most suitable forvisualization of the sweat gland apparatus.10 Theresults, with absence of the acrosyringium in all speci-mens from involved PPP skin and preserved ducts in allcontrol subjects and in the patients with tylotic eczema,indicate that there is a destruction of the intraepidermalduct in PPP. One possible interpretation of our results isthat that there may be an inflammatory process in PPPthat involves and destroys the intraepidermal sweatduct with participation not only of lymphocytes andneutrophils but also, in a hitherto unknown way, ofeosinophils and mast cells. The disappearance of theepidermal part of the duct may be a non-specific feature,but the preserved eccrine ducts in our patients withtylotic eczema as well as in patients with dyshidroticeczema17 may indicate that the absence of a normalepidermal duct structure in PPP is of pathogeneticrelevance and that the pustule formation involves theacrosyringium. It is not yet known if similar changes arepresent in palmar/plantar psoriasis but it has beenreported that acrosyringial keratinocytes in psoriasis

396 M-O.ERIKSSON et al.

q 1998 British Association of Dermatologists, British Journal of Dermatology, 138, 390–398

Page 8: Palmoplantar pustulosis: a clinical and immunohistological study

do not display an abnormal differentiation.18 Thereare, however, several reports on reduced sweating inpsoriasis plaques.19,20

In this study it was further confirmed that thyroiddisease is common in PPP patients. Furthermorelaboratory abnormalities (TSH, T4, antibodies to thyro-globulin and thyroperoxidase) were common in patientsnot previously known to have thyroid disease. Howautoimmune thyroid disease can be linked to PPP isnot known. It is noteworthy, however, that there is ahomology between thyroglobulin and acetylcholinester-ase and between thyroid peroxidase and myeloperoxi-dase.21 There are also other associations betweenthyroid hormones and keratinocytes which may berelevant with regard to PPP. For example, thyroidhormones bind to specific nuclear thyroid hormonereceptors, which participate in the regulation of keratingenes.22

The presence of increased IgA antibodies to gliadin,increased serum IgA and decreased serum IgM has notpreviously been reported in PPP. However, we haveobserved a few PPP patients with previously unknowninflammatory changes in the duodenal mucosa similarto those in dermatitis herpetiformis and in psoriasis.23

One of these patients had a disabling PPP, which sub-sequently healed on a gluten-free diet.

PPP patients have been reported to smoke more thanhealthy subjects. In our study all 56 of smokers orex-smokers of our 59 patients were smokers at theonset of the PPP disease. Similar figures were reportedby Rosen et al.1 This may indicate that nicotine mayprecipitate PPP. How this can be linked to the finding ofabnormal intraepidermal sweat ducts is not known, butone possibility is that the effect of nicotine on thefunction of the sweat gland and its duct might be ofimportance. Nicotine can also influence keratinocytefunction, as it has been reported to induce cornificationof keratinocytes in vitro.24 Also, keratinocytes express anicotinic acetylcholine receptor regulating cell adhesionand motility.25 The possible relation between the destruc-tion of the acrosyringium, the psoriasiform inflam-mation, autoimmune thyroid disease and smoking willbe further investigated.

Acknowledgments

This project was supported by grants from the SwedishPsoriasis Association, the Edvard Welander Foundationand the Swedish Foundation for Health Care Sciencesand Allergy Research.

References1 Rosen K. Pustulosis palmo-plantaris and chronic hand dermatitis.

Treatment, epidermal Langerhans cells and association withthyroid disease. Acta Derm Venereol (Stockh) 1988; 137 Supple:1–52.

2 Akiyama T, Seishima M, Watanabe H, Nakatani A et al. Therelationships of onset and exacerbation of pustulosis palmaris etplantaris to smoking and focal infections. J Dermatol 1995; 22:930–4.

3 Kubeyinje EP, Belagavi CS. Risk factors for palmo-plantarpustulosis in a developing country. East Afr Med J 1997; 74:54–5.

4 Lundin A, Hakansson L, Michaelsson G, Venge P. Neutrophillocomotion and serum chemotactic and chemokinetic activitiesin pustulosis palmo-plantaris compared with psoriasis. Arch Der-matol Res 1987; 279: 385–91.

5 Lundin A, Fredens K, Michaelsson G, Venge P. The eosinophilgranulocyte in psoriasis. Br J Dermatol 1990; 122: 181–93.

6 Michaelsson G, Kraaz W, Gerden B et al. Patients with psoriasisoften have increased serum levels of IgA antibodies to gliadin. Br JDermatol 1993; 129: 667–73.

7 Michaelsson G, Kraaz W, Gerden B et al. Patients with psoriasishave elevated levels of serum eosinophil cationic protein andincreased numbers of EG2 positive eosinophils in the duodenalstroma. Br J Dermatol 1996; 135: 371–8.

8 Axelsson L, Bergenfeldt M, Ohlsson K. Studies of the release andturnover of a human neutrophil lipocalin. Scand J Clin Lab Invest1995; 55: 577–88.

9 Michaelsson G, Kraaz W, Hagforsen E et al. Psoriasis patients havehighly increased numbers of tryptase-positive mast cells in theduodenal stroma. Br J Dermatol 1997; 136: 866–70.

10 Watanabe S, Ichikawa E, Takanashi S, Takanashi H. Immunohis-tochemical localization of cytokeratins in normal eccrine glands,with monoclonal antibodies in routinely processed, formalin-fixed,paraffin-embedded sections. J Am Acad Dermatol 1993; 28: 203–12.

11 Uhehara M, Ofuji S. The morphogenesis of pustulosis palmaris etplantaris. Arch Dermatol 1974; 109: 518–20.

12 Baggiolini M, Walz A, Kunkel SL. Neutrophil-activating interleu-kin 8, a novel cytokine that activates neutrophils. J Clin Invest1989; 84: 1045–9.

13 Burrows LJ, Piper PJ, Lindley ID, Westwick J. Intraperitonealinjection of human recombinant neutrophil-activating factor/interleukin 8 produces a T cell and eosinophil infiltrate in theguinea pig lung. Effect of PAF antagonist WEB 2086. Ann NY AcadSci 1991; 629: 422–4.

14 Anttila Heli SI, Reitamo S, Erkko P et al. Interleukin-8 immuno-reactivity in the skin of healthy subjects and patients withpalmoplantar pustulosis and psoriasis. J Invest Dermatol 1992;98: 96–101.

15 Moller A, Lippert U, Lessman D et al. Human mast cells produceIL-8. J Immunol 1993; 15: 3261–6.

16 Cox A. Mast cells in psoriasis. In Psoriasis. Proceedings of the SecondInternational Symposium, 1976 (Farber E, Cox A, eds). New York:Yorke Medical Books, 1977; 36–43.

17 Kutzner H, Wurzel RM, Wolff HH. Are acrosyringia involved in thepathogenesis of ‘dyshidrosis’? Am J Dermatopathol 1986; 8: 109–16.

18 Wollina U, Schaarschmidt H, Knopf B, Hipler C. Keratinozytendes Acrosyringium-keine Differenzierungsstorung bei Psoriasisvulgaris. Dermatol Monatsschr 1990; 176: 151–5.

19 Juhlin L. Effect of treatment on sweat retention in psoriasis. ActaDerm Venereol 1967; 47: 98–101.

PALMOPLANTAR PUSTULOSIS 397

q 1998 British Association of Dermatologists, British Journal of Dermatology, 138, 390–398

Page 9: Palmoplantar pustulosis: a clinical and immunohistological study

20 Johnson C, Shuster S. Eccrine sweating in psoriasis. Br J Dermatol1969; 81: 119–24.

21 Mcgregor AM. The thyroid gland and disorders of thyroid function.In: Oxford Textbook of Medicine (Weatherall DJ, Ledingham JGG,Warrell DA, eds). Oxford: Oxford University Press, 1996; 1603–18.

22 Tomic-Canic M, Day D, Samuels H et al. Novel regulation of keratingene expression by thyroid hormone and retinoid receptors. J BiolChem 1996; 271: 1416–23.

23 Michaelsson G, Kraaz W, Hagforsen E et al. Increased lymphocyte

infiltration in duodenal mucosa from patients with psoriasis andserum IgA antibodies to gliadin. Br J Dermatol 1995; 133: 896–904.

24 Theilig C, Bernd D, Ramirez-Bosca A et al. Reactions of humankeratinocytes in vitro after application of nicotine. Skin Pharmacol1994; 6: 307–15.

25 Grando SA, Horton RM, Pereira EF et al. A nicotinic acetylcholinereceptor regulating cell adhesion and motility is expressed inhuman keratinocytes. J Invest Dermatol 1995; 105: 774–81.

398 M-O.ERIKSSON et al.

q 1998 British Association of Dermatologists, British Journal of Dermatology, 138, 390–398