The probability that Parkinson’s disease (PD) would be a mendelian disorder is higher in patients with age onset younger than 50. Three different genes, PRKN2 (OMIM # 602544), PINK1 (OMIM #608309) and PARK7 (OMIM #602533), have been identified to cause autosomal recessive young onset PD forms [1] - [3] . PRKN2 encodes a protein of 465 amino acids named Parkin which functions as an E3-Ubiquitin protein ligase [4] [5] . The clinical symptoms of PRKN2 patients are variable, but similar to those found in idiopathic PD. Its phenotype is usually characterized by onset before the age of 40 years (may range from 7 to 72 years), good response to levodopa therapy, dyskinesias under low doses of levodopa and slow progression. Additional features such as dystonia and absence of Lewy bodies (LBs) in the substantia nigra [1] [6] [7] are also frequent. Although the absence of LBs in the substantia nigra seems to be frequent in PRKN2 mutation carriers, three PRKN2 cases have been reported to show LB pathology [8] [9] .

LRRK2/dardarin mutations are now a well established cause of typical PD [10] . In our geographical region, the Basque Country (Spain), the prevalence of the R1441G mutation is relatively high; in fact, a PD familial patient of Basque origin has a probability of 46% to carry this mutation [11] . The prevalence of G2019S in this region is similar to other European countries [11] .

Here, we analysed the prevalence of PRKN2 and both R1441G and G2019S LRRK2/dardarin mutations in patients with PD with an age at onset younger than 50 years of a geographical area (Gipuzkoa) of the Basque Country. In addition, we inspected the phenotypic differences between PD patients carrying PRKN2 mutations and those who were non carriers to establish a range of clinical manifestations that allowed us to make a more accurate molecular analysis in our patients with early onset parkinsonism (EOPD)

Thirteen out of twenty seven patients with EOPD studied carried PRKN2 mutations (further details are described in Table 1). In addition, the R1441G LRRK2 mutation was identified in a PD patient carrying as well a het- erozygous deletion involving exons 3 and 12 of the PRKN2 gene. The G2019S mutation was present in four out of fourteen patients who did not show any pathogenic PRKN2 variant (further details are described in Table 2).

Sex: M: male, F: female; Family history: Mt: mother, Un: uncle, B: brother; Symptoms: T: tremor, LL: low limb, Uni: unilateral, A: akinesia, Bi: bilateral, UL: upper limb, G: gait disturbance; L-dopa response: G: good, R: regular, P: poor, NT: not tolerated; Progression: VB: very benign (−), B: benign (+), R: regular (++), M: malignant (+++); DysK: dyskinesias; Fluct: fluctuations.

Sex: M: man, F: woman; Family history: Mt: mother, Un: uncle, B: brother, F: Father, GF: grandfather; Symptoms: T: tremor, LL: low limb, Uni: Unilateral, A: akynesia, Bi: bilaterall, UL: upper limb; G: gait disturbance; R: rigidity, Dy: dystonia; L-dopa response: G: good, R: regular, P: poor, NT: not tolerated, Und: undeterminated; Progression: VB: very beningn (−), B: benign (+), R: regular (++), M: malign (+++), Und: undeterminated; DysK: dyskinesias; Fluct: fluctuations.

In the Province of Gipuzkoa (The Basque Country, Spain), where the prevalence and incidence of PD is similar to other European countries [17] , the prevalence of the p.Arg1441Gly/LRRK2 mutation is relatively high [11] . In this study we have also found a high prevalence of PRKN2 mutations. Therefore, one may speculate that other mutations may be more prevalent in different geographic areas. In this work we present prevalence of PRKN2 and LRRK2 mutations in a cohort of EOPD in a region with a population which is possibly unique and different from others.

An extensive review of PRKN2 mutations identified the most common mutations as deletions of exon 4 (7.38%), deletions of exon 3 (6.06%), deletions of exons 3 and 4 (6.06%), a point mutation in exon 7 (c.924C > T, 10.02%) and a single base pair deletion in exon 2 (c.255-256delA, 4.48%) [18] . Four of these five common mutations occurred in our sample, confirming that the hot spots for small mutations may be concentrated in exon 2 and for break points in introns 2 through 4. In our sample, the most common mutation found was c.255-256delA. The frequency of homozygous carriers for c.255-256delA has been estimated at 1/40.000 in the Spanish population [19] , supporting the hypothesis of a possible Spanish founder for this mutation.

The presence of family history is relatively high in our sample analyzed (66.67%); indeed, 44.45% showed an autosomal recessive pattern of inheritance with affected siblings carrying the same genetic defect. All autosomal recessive patients carried PRKN2 mutations, compiling a total of 29.6% of the cases here described. This proportion is similar to those reported in other studies [20] - [22] . In addition, the patient who carries heterozygous PRKN2 deletions involving exons 3 and 12, and the R1441G mutation of LRRK2, had a family history compatible with autosomal dominant inheritance (two unaffected brothers and an affected parent). An additional PRKN2 mutation carrier had also an affected uncle.

With respect to the cases without PRKN2 mutations, 6 familial individuals presented with strong genetic background compatible with an autosomal dominant pattern of inheritance.

The phenotypic characteristics found in the cases reporting PRKN2 mutations resemble to those described in most of the genetic studies to date reported [20] - [23] . The PRKN2-associated phenotype is well defined and broadly similar; the main clinical manifestations are the young age at onset, and the benign disease progression. The majority of the cases began with tremor which can be unilateral and/or progress to be bilateral and occasionally accompanied with postural tremor. Additionally, some of the cases presenting with akinesia or gait disturbances have a more severe disease progression than those initially presenting with tremor, being similar to that found in PRKN2 mutation non-carriers who showed dystonia or akinesia as initial symptom. None of our patients have dystonia, diurnal fluctuations, exercise induced dystonia, psychosis or peripheral neuropathy described elsewhere [23] .

In this study the evolution in patients with PRKN2 mutations is more benign. In 5 patients, we considered the evolution very benign, presenting only mild parkinsonian signs in exploration under small doses of antiparkinsonian medication, many years after the onset of symptoms. Phenotype-genotype dissociations were observed in 2 sisters presenting with very different clinical evolution, suggesting that other genetic or environmental factors and behaviours may possibly justify these differences. The PD with double mutation (PRKN2/LRRK2) had a similar age at onset (44 years) to the other PRKN2 mutation carriers; suggesting that to carry pathological mutations in both genes does not have an effect on the disease features.

An important issue is to reflect on the value of the molecular study of the PRKN2 gene in EOPD. In our series, although the age of onset of symptoms did not differ statistically between both groups (36.7 in PRKN2 carriers versus 41.7 in the PRKN2 non carriers), probably because of the sample size, PD patients that present with the disease after age 45 are less likely to have a PRKN2 mutation. Autosomal recessive patients with an age at onset lower than 45 years and showing a typical idiopathic PD are highly susceptible to carry PRKN2 mutations.

Given the lack of screening for PARK7 and PINK-1 mutations and PRKN2 promoter variations in our negative PRKN2 mutation cases, we cannot exclude that in these cases PD may be caused by alterations in one of these genes. Other mutations in LRRK2 are uncommon and did not be studied. In our study we observed that PRKN2 mutations are more frequent in early onset familial cases than in early-onset sporadic cases (frequency of 0.26).

We recognize in this work the limitation that supouse of having looked for only the p.Arg1441Gly and p.Gly2019Ser mutations in the LRRK2 gene, and mutations in PKRN2 gene. It is true that the study would have been more consistent if the number of patients studied had been greater, and if we had included other pathogenic mutations in the LRRK2 gene and mutations in other genes involved in recessive forms of PD as PINK1 and PARK7.

In our opinion, with the limitations of this small sample, in our region knowing whether someone may or may not carry PRKN2 mutations have some advantages, such as warning about the possibility to be a carrier and helping in an accurate genetic counseling, and somehow in the prognosis of the disease; indeed, the therapeutic strategy might be different.

Forty eight percent of the 27 PD patients with age of onset before 50 were PRKN2 mutation carriers. The c.255- 256DelA mutation was the most frequent, followed by a deletion involving exons 3 and 4. A deletion involving 3 and 12 of the PRKN2 gene and R1441G LRRK2 mutation was found together in one PD patient.

Four out of fourteen PRKN2 negative patients carried the pG2019S mutation of the LRRK2 gene.

Both PRKN2 mutation carriers and non-carriers presented frequently with family history (10 PRKN2 mutation carriers and 8 PRKN2 non-carriers); in fact, five patients without a known gene mutation had a first degree relative affected, suggesting another monogenic disease.

PRKN2 carriers presented with a younger age at onset (36.7 vs. 41.7) and more benign disease progression. Indeed, those PD patients younger than forty who initially presented with unilateral tremor became shortly bilateral. Relatively, symmetric parkinsonism and slow disease progression carried more frequently PRKN2 mutations than patients with unilateral akinetic rigid parkinsonism and age at onset later than 40 years. As expected in a recessive disease, PRKN2 patients present more often with affected siblings and unaffected parents.

The G2019S LRRK2 mutation, less prevalent than R1441G in our area, may be also a frequent cause of PD in EOPD (4 patients).

We thank the patients and their families for taking part of this study. We wish to thank the neurologist who to assist in recruitment, and Dra. Connie Marras for carefully reading the manuscript.

Centro de Investigación Biomédica en Red sobre Enfermedades Neurodegenerativas (CIBERNED), Instituto de Salud Carlos III (PI070660), Eusko Jaurlaritza (Osasun Saila) (2006111003), Gipuzkoako Foru Aldundia, Fundación de Investigación Médica Mutua Madrileña, Ilundain Fundazioa.

José F Martí Massó was responsible for study coordination, data collection, and manuscript preparation. Javier Ruiz-Martínez. Alberto Bergareche and Adolfo López de Munain assisted with data collection and manuscript preparation. Coro Paisan-Ruiz, Ana Gorostidi, Ainhoa Alzualde, and Jordi Perez-Tur were responsible for molecular studies and manuscript preparation.

All authors reviewed the manuscript.

This work was approved by the ethic committee of Hospital Donostia.