Peripheral precocious puberty in Li–Fraumeni syndrome: a case report and literature review of pure androgen-secreting adrenocortical tumors

Established facts

Novel insights

Peripheral precocious puberty (PPP), contrary to central precocious puberty (CPP), is caused by the autonomous secretion of androgens or human chorionic gonadotropin (hCG). Differentiating CPP from PPP in males is relatively straightforward and based on clinical and hormonal findings. In CPP, boys have testicular growth and elevated gonadotrophin levels, whereas in PPP, testicular volume is prepubertal and gonadotrophins are suppressed. However, diagnosing the cause of PPP can be more challenging, and includes neoplasms such as Leydig cell tumors, hCG-producing germ cell tumors, and androgen-secreting adrenocortical tumors (ACT) [1].

ACT are very rare in children; they can be benign or malignant, and their reported annual incidence is 0.2–0.3 new cases per million per year [2]. Contrary to adults, most ACT are functional in children and secrete cortisol, often together with androgens [3]. They are usually sporadic and benign but sometimes associated with tumor predisposition syndromes, such as Beckwith–Wiedemann syndrome and Li–Fraumeni syndrome (LFS) [4]. However, pure androgen-secreting adrenocortical tumors (PASACT) are ultra rare, and most reports describe solitary cases. We report a 2.5-year-old boy who presented with isosexual PPP due to a PASACT as the first expression of LFS, complicated by apparent tumor-related arterial hypertension. This case illustrates the differential diagnosis of PPP in young boys and highlights the importance of performing histopathological and genetic analysis upon diagnosing PASACT. We also reviewed previous reports of children with PASACT to identify common characteristics and potential clinical, imaging, or biochemical features that could allow preoperative differentiation between adrenocortical adenomas and carcinomas.

A 2.5-year-old boy presented with a 6-month history of penile enlargement, pubic hair, frequent erections, and rapid linear growth. Delta height was + 1.19 standard deviation score (SDS) in the past 8 months. Previous medical and family history was unremarkable, and exposure to exogenous testosterone was excluded. On physical examination, his weight was 19.1 kg (+ 2.5 SD), height 101 cm (+ 1.7 SD), systolic blood pressure 138 mmHg (> 99th percentile), and diastolic 75 mmHg (95th–99th percentile). Tanner stage was A1 P4 G3, and penile length was 6.8 cm. He had symmetrical prepubertal testes (3 ml bilateral) without palpable abnormalities. Skin and oral examination were normal. Repeated blood pressure readings in the following weeks were all above the 95th percentile. His bone age was 3.5 years, as per the Greulich–Pyle scale. Serum gonadotrophin levels were prepubertal, while testosterone (9.9 µg/L, ref. < 0.12 µg/L), androstenedione (2365 ng/l, ref. 100–900 ng/L), and dehydroepiandrosterone-sulfate (DHEA-S) levels (1.49 mg/L, ref. < 0.02–0.15 mg/L) were elevated, approaching the adult reference range. Serum electrolytes, 17-OH-progesterone, adrenocorticotropic hormone (ACTH), renin, aldosterone, and cortisol levels were normal, as were repetitive 24-hour urinary free cortisol excretion and urinary catecholamine levels (Table 1). Testicular ultrasound was normal. Abdominal ultrasound showed a hypoechogenic ovoid nodule in the left adrenal (Fig. 1A), which on magnetic resonance imaging (MRI) was well defined (26 × 23 × 31 mm) and hyperintense on T2 weighted sequences. Contrast accumulation was limited and homogeneous, and there was no evidence of invasion into adjacent organs or vessels (Fig. 1B). Chest X-ray and liver ultrasound imaging were normal. We decided on laparoscopic unilateral adrenalectomy on the basis of the hormonal findings and imaging features. Two days after surgery, blood pressure normalized, and normal prepubertal levels of testosterone, androstenedione, and DHEA-S were found.

Macroscopically, the tumor was well defined and surrounded by a thin capsule (Fig. 2A). Histologically it contained areas of nuclear pleomorphism but no necrosis (Fig. 2B, C). There were no signs of capsular or vascular invasion. The tumor stained positive for inhibin (Fig. 2D) and Melan-A (Fig. 2E). It was classified as a cortical neoplasm with a favorable prognosis on the basis of (immuno)histological characteristics (Wieneke criteria, score: 0): tumor weight < 200 g, no capsular invasion or tumor necrosis, mitotic activity < 15/20 high-power field (HPF), Ki67 index < 15% (Fig. 2F), and strong and diffuse P53 expression (3 + intensity, 100% of positive cells) (Fig. 2G). The comparative genomic hybridization of the tumor revealed a complex karyotype with multiple trisomies (Chr 1, 5, 7, 8, 9, 14, 16, and 19), monosomies (Chr 2, 3, 4, 10, 11, 17, and 21), and loss of chromosomes X and Y.

Massive parallel sequencing was performed on the tumor, analyzing a total of 165 cancer-related genes (additional file 1: Table S1), which identified a pathogenic TP53 gene variant NM_0000546.5(TP53):c.473G > A, p.(Arg158His) (conform ClinVar database https://​www.​ncbi.​nlm.​nih.​gov/​clinvar), at an allelic frequency of 71%. Targeted Sanger sequencing on DNA from a whole blood sample subsequently confirmed the heterozygous germline presence of this pathogenic TP53-variant, confirming the diagnosis of LFS (see also Fig. 3A–C). Familial screening of the proband’s parents and brother was negative, indicating de novo occurrence of the variant. Two years after the adrenalectomy, the patient is in good health, and whole-body MRI cancer screening remains negative for abnormalities. Of note, there was no regression of the penis development or pubic hair.

We conducted a literature search in Pubmed using the query “androgen AND adrenal AND pediatric AND (carcinoma OR adenoma)”. From the 76 retrieved articles, we selected 7 reports with detailed clinical and histological data. Finally, we retrieved 16 detailed case reports using citation searching. We summarized their clinical and histological findings, and when available, the results of genetic testing in Table 2 and compared them with our patient [5‐19].

The median age at diagnosis was 4 years (range: 2 months–18 years), and there was a female sex bias (7 males versus 9 females). All patients had clinical signs of androgen excess, such as pubic hair before 8 years of age or voice deepening in five cases. In 5/7 male cases, patients had rapid or early penile growth, while clitoral hypertrophy was reported in just 2/9 female cases. In 6/16 cases, there was facial acne, and a single subject (a 3-year-old boy) had precocious axillary hair development. In 2/3 female patients older than 12 years (12 years 10 months, 15 years, and 18 years), there was primary amenorrhea, but one adolescent girl had regular menses.

Testosterone levels were elevated in all patients, and all investigated children had normal serum cortisol, 24-hour urinary free cortisol, and plasma adrenocorticotropic hormone levels (Table 2). The commonly used imaging modalities were abdominal ultrasound, CT, or MRI. Tumor laterality was symmetrical (9 right versus 8 left adrenal tumors). However, in one case the intraadrenal adenoma on the right side was accompanied by an ectopic adrenal adenoma in the left renal hilus. The median tumor size was 3.05 cm in diameter (< 1–11 cm). Eight tumors were histologically classified as adenoma, five as carcinoma, and four were not specified. All adenomas were smaller than 5 cm except for two cases. Cases of adenoma or carcinoma did not differ in age at presentation, sex distribution, presenting symptoms, tumor size, or laterality. Vascular or soft tissue invasion was absent in both adenomas and carcinomas in all 17 reported tumors in the 16 reported cases. None of the reported cases presented metastatic disease. Genetic testing for germline TP53 gene variants was performed in 4/16 patients. Here, to the best of our knowledge, we report for the first time an underlying pathogenic TP53 variant in childhood PASACT.

Penile growth without testicular enlargement in boys under 9 years of age is a typical sign of PPP. The differential diagnosis of PPP in boys is broad and includes exposure to exogenous testosterone, testotoxicosis, and non-classical congenital adrenal hyperplasia, but also neoplastic causes, including Leydig cell tumors, hCG-producing germ cell tumors, and androgen-secreting ACT [20]. Therefore, initial screening should include bone age assessment and measurement of gonadotropins, testosterone, hCG, DHEA-S, and 17-OH-progesterone levels. In the case of mainly elevated serum testosterone and androstenedione concentrations—as in this report—either a testicular or adrenal tumor should be sought. Indeed, in this case, we tentatively diagnosed a PASACT on the basis of clinical (well-defined nodule in the left adrenal gland with lack of invasion of adjacent tissue on MRI imaging and a normal testicular sonogram) and biochemical (high androgens and normal renin-aldosterone blood serum levels and normal urinary cortisol excretion) features.

In case of isolated hypercorticism and negative adrenal imaging, further workup with a dexamethasone test is warranted. However, in this case of PPP there were no signs of hypercorticism and imaging revealed the adrenal tumor. Therefore, no dexamethasone suppression test was carried out.

PASACT is a rare neoplastic cause of PPP that is either benign or malignant and most often presents under the age of 5 years [21, 22]. The initial treatment step for PASACT is surgical tumor resection, which we performed by unilateral laparoscopic adrenalectomy. However, the current literature shows that tumor size or soft tissue invasion cannot differentiate adenomas from carcinomas reliably. This implies that clinical signs and imaging cannot differentiate benign from malignant PASACTs. However, after surgery, histological tumor features such as higher mitotic rate, higher percentage of necrosis, and larger tumor size are more associated with carcinomas [23].

In the present case, the histology (low mitotic rate, no necrosis) and the normalization of androgen levels after surgery confirmed the diagnosis of a pure androgen-secreting adrenal adenoma. However, the histology of ACT in children is challenging, as the classical Weiss criteria used in adults often misclassify childhood ACT as carcinomas [24]. Therefore, a distinct pediatric grading system, the Wieneke criteria, should be combined with the Ki67 index and p53 status for childhood ACT [25‐27]. Here, in the present case, both the Wieneke criteria (score: 0) and the Ki67 index (< 15%) indicated an adrenal adenoma. However, p53 status (100% 3 + nuclear p53 expression) suggested a TP53-mutation. Among the previously reported PASACT cases, only five mentioned the Ki67 index or the Weiss criteria, and none reported using the appropriate Wieneke criteria.

Both benign and malignant ACTs are reported in genetic tumor syndromes, such as Beckwith–Wiedemann syndrome, Carney complex, and LFS. In LFS, adrenal carcinomas have mainly been reported. In our case, there were no clinical signs of a tumor predisposition syndrome, and the family history was negative for tumors. Germline pathogenic variants in TP53 are found in approximately 45–80% of children with adrenocortical carcinomas and a negative family history of LFS [28, 29]. Moreover, in up to 20% of LFS families, the TP53 variants arose de novo. Hence, with its variable expression and penetrance, families with LFS are prone to be missed [30]. Nonetheless, germline TP53 variants predisposing to LFS are found in more than half of children with ACT, but the majority were cortisol-secreting [31]. The tumor was characterized as an adenoma; however, it is unclear whether adrenocortical carcinomas arise as such or if they can originate from benign lesions such as hyperplasias or adenomas [32]. Therefore, we performed a TP53 gene analysis in the proband. The specific TP53 missense variant [NM_000546.5(TP53):c.473G > A] found in the presented case has been reported multiple times, and the resulting p.R158 substitution hits a functional part of the TP53 DNA-binding domain (Fig. 3C). Varley et al. were one of the first to describe this germline variant in 3/14 subjects in a study on childhood ACTs [28]. In a larger study regarding TP53 gene variant carriers in LFS, Bougeard et al. reported that out of the 11 carriers of the p.Arg158His variant (from 8 different families), 4 were positive for a history of childhood adrenocortical carcinoma (ACC) [33]. Furthermore, this variant was the most identified germline alteration in 34 children with ACC [33].

It is important to identify the underlying type of TP53 variant for clinical follow-up. Missense variants, including p.Arg158His, differ from null variants as patients with missense variants present approximately 9 years earlier than those with other TP53 variants, including null mutations, suggesting an additional oncogenic effect in TP53 missense variants [29]. Correspondingly, in a simplified p53 functional assay to quantify the induction of p53 transcriptional activity in response to DNA damage after exposure to doxorubicin, the mean functionality score in the p.Arg158His variant was about 50% of the mean score of the control group. This 50% reduction in functionality was like the remaining function in TP53 null variants, indicating a dominant negative effect [34]. Thus, intensive tumor surveillance with annual total body and brain MRI is needed in these patients, and exposure to ionizing radiation should be avoided. Furthermore, long-term endocrine follow-up is indicated as gonadotropin-dependent pubertal disorders are common in children with a history of virilizing ACT [35].

The arterial hypertension in the present case posed an additional diagnostic challenge, since both an excess of circulating secreted adrenal hormones and their metabolites or renal artery compression by a large ACT, with concomitant secondary hyperaldosteronism, were possible. Urinary cortisol excretion, renin–aldosterone state, and normal urinary catecholamines were normal, and renal artery compression was not detected. Furthermore, the rapidly normalized blood pressure after tumor resection strongly suggests a tumor-driven mechanism. Androgens have a blood-pressure-elevating effect. This has been studied in both animal models and humans. Blood pressure decreased after castrating male rats and increased after administration of testosterone both in normotensive female and in normotensive male rats [36]. Similarly, in transgender population studies, testosterone increased blood pressure [37]. An association of male sex steroids with the reduction of pressure natriuresis has been proposed previously as a contributing cause of hypertension [15]. In our review of childhood PASACT cases, only one patient had arterial hypertension, which normalized 2 months after surgery [15]. In this case, the levels of 11-deoxycorticosterone, a potent mineralocorticoid, were normal. However, mineralocorticoid receptor stimulation and subsequent hypertension by adrenal steroids that are not measured routinely cannot be excluded and might be involved in some PASACT cases [38, 39].

Androgen-secreting tumors should be part of differential diagnosis in children with PPP. Clinical, biochemical, and imaging features are important for diagnosing an androgen-secreting ACT, but cannot differentiate benign from malignant tumors. Molecular genetic testing is important to ascertain whether an ACT is related to a specific genetic tumor predisposition syndrome, including LFS, even when histologically considered as having a favorable prognosis. Arterial hypertension can be associated with pure androgen-secreting adrenocortical adenomas in children.