Characterization of Immunodeficiency in a Patient With Growth Hormone Insensitivity Secondary to a Novel STAT5b Gene Mutation

CASE REPORTS

The patient, a 16-year-old girl, was the first of 2 children of nonconsanguineous parents. She was born with a birth weight of 2500 g. She has had severe growth failure since the first year of life. She was first referred to another institution at the age of 6 years, at which time her height was 83.6 cm (SD score; [SDS]: −6.17), her weight was 9800 g (relative weight: 89%), and her bone age was 2.5 years (Fig 1).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1

Our patient's growth chart. Height (in centimeters) and bone age (in years) are depicted by the • and × symbols, respectively. Horizontal lines indicate the corresponding chronological and bone ages. Severe growth failure and delayed bone age were already present at 6 years of age.

Since her first months of life she presented generalized eczema and recurrent severe infections of the skin and respiratory tract that required multiple hospitalizations. She also presented recurrent episodes of pneumonia, for which she was hospitalized and treated with intravenous antibiotics. No microbiologic isolations were reported.

She had been suffering from chronic diarrhea since she was 2 months old. Gastroenterological and laboratory evaluations ruled out cystic fibrosis and celiac disease. Ascaris lumbricoides and Enterobius vermicularis were found at different times; to our knowledge, no other organisms were isolated. At the age of 4 years she had prolonged varicella complicated with bacterial infection. Since the age of 10 years, she has had recurrent herpes zoster keratitis and uveitis of the left eye, with progressive loss of visual acuity.

No family history of immunodeficiency or growth failure was reported. However, short stature was observed in her mother. Paternal and maternal heights were 160 cm (SDS: −2.2) and 142.3 cm (SDS: −3.3), respectively. Her 10-year-old brother had normal height. At 16 years of age she was hospitalized because of respiratory distress and was referred to our immunology and endocrinology units. Clinical features of congenital GH deficiency were observed: prominent forehead, saddle nose, and high-pitched voice, with no signs of pubertal onset. She had a persistently low growth rate and severe growth failure (height: 102.7 cm [SDS: −9.91]; weight: 12 kg [SDS: −6.7]), and her bone age was severely delayed (7 years old). Basal serum triiodothyronine, thyroxine, thyrotropin, cortisol, luteinizing hormone, and follicle-stimulating hormone were normal for age and pubertal development. However, here serum prolactin level was persistently high in 3 different samples drawn with an interval time of 1 month (168, 102.3, and 143 ng/mL).

She had evidence of chronic lung disease with signs of chronic hypoxemia (clubbing). Active herpes keratitis and severe eczema were also present. Pharyngeal swabs and blood and sputum cultures were negative for virus, fungus, bacteria, and mycobacteria.

Serology analysis showed negative results for HIV, cytomegalovirus, and Epstein-Barr virus (EBV) immunoglobulin (Ig)M but positive results for cytomegalovirus and EBV IgG. The thorax computed tomography scan showed a ground-glass appearance with generalized micronodular images and bronchiectasis in the left lower lobe. Pulmonary-function tests showed mixed, restrictive and obstructive ventilative insufficiency of moderate degree. She was treated empirically with antibiotics without any response.

Lymphoid interstitial pneumonia was suspected; a lung biopsy was indicated, but the parents refused the test. Since then, the patient's pulmonary function has worsened, and permanent oxygen has been required. Because of the long history of recurrent infections, chronic lung disease, and severe growth retardation, immunodeficiency and GH deficiency or insensitivity was suspected.

The patient was lost to follow-up. Treatment with high doses of prednisone was started in another institution and, after 3 months, a partial improvement of the respiratory condition was observed (E.M. Rivas, verbal communication, 2005).

Our study was approved by the institutional review board of the Garrahan Pediatric Hospital. Informed consent, as well as assent, were obtained from the parents and patient, respectively.

METHODS

RESULTS

Immunologic Studies

Moderate lymphopenia and hypergammaglobulinemia, mainly of IgG and IgE isotypes, was observed (Table 1). Serologic evaluations showed positive IgG antibodies for varicella herpesvirus, hepatitis A virus, EBV, cytomegalovirus, and herpes simplex virus I and II. Serum antibodies specific to recall antigens (tetanus toxoid and Pneumococcus) were within the reference range (Table 1), showing integrity of the antibody function. Complement proteins were below the reference levels, and a high titer (1/1000) of antinuclear antibody was detected.

View this table:

• View inline
• View popup

TABLE 1

Immunological Parameters

The patient had moderate T-cell lymphopenia, with numbers of CD4 and CD8 T cells in the lower limit and below the reference range, respectively, but within the reference CD4/CD8 ratio. B cells were within the reference range. It is notable that an extreme reduction in percentages and in absolute counts of NK and γδ T cells was found (Table 1).

T lymphocytes showed a phenotype of chronically activated cells: skewed CD45RA/CD45RO ratio, high percentages of HLA-DR, CD95, and CD25 (Table 1 and Fig 2 A–F), and down-modulation of CD62L (Fig 2F).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2

Our patient's T-cell–activation profile. A, Expressed in 100% of T cells; B–F, 100% of CD4 T cells. Analysis of Tregs was performed by selecting the CD4⁺CD25^high T-cell population indicated by a circle (in C). FITC indicates fluorescein isothiocyanate; PerCP, perydin clorophyl protein; PE, phycoerythrin.

Analysis of the CD4⁺CD25⁺ Treg population was based on the definition of Tregs by Baecher et al.²⁵ Despite the fact that T cells were hyperactivated, a minor population with high expression of CD25 could be detected that represented only 1.16% ± 0.36% (n = 3) of the total CD4 T cells (Fig 2C). According to our age-matched reference values of 4.04% ± 0.26% (n = 7), they were clearly diminished (P < .01). These cells constitutively expressed cytotoxic T-lymphocyte–associated protein 4 (CTLA4) and were CD62L^+high, indicating that, phenotypically, they were truly Tregs²⁵ (data not shown).

In vitro T-cell proliferation is shown in Table 2. Proliferation to mitogens (phytohemagglutinin, S enterotoxin B, phorbol myristate acetate/ionomycine, and anti-CD3) was consistently low and could not be restored to reference values despite rIL-2 stimulation. After natural exposure or vaccination, absence of proliferative response to recall antigens (tetanus toxoid, purified protein derivative, and Candida) was observed (Table 2). These results support the concept of a general functional T-cell defect besides the defect in the IL-2 signaling pathway.

View this table:

• View inline
• View popup

TABLE 2

T-Cell Proliferation in Vitro: Incorporation of ³[H]thymidine in cpm (Mean of Triplicates)

The presence of high T-cell activation, hypergammaglobulinemia, antinuclear antibodies, and diminution of CD4⁺CD25⁺ Tregs suggests a defect in the peripheral homeostasis mechanisms.

To analyze the IL-2 signaling pathway, STAT5 phosphorylation under IL-2 stimulation was evaluated by FC.

In the patient's lymphocytes, as shown in Fig 3A, STAT5 protein phosphorylation in response to high concentrations of rIL-2 was reduced to less than half of the control MCFI value (206 vs 462 ± 51.8 [n = 3]).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3

Reduced rIL-2-induced STAT5 phosphorylation and total STAT5 protein expression in patient's PBMCs. A, Phospho-STAT5 expression in IL-2-stimulated cells (shaded histogram) overlaid with histograms of phospho-STAT5 in unstimulated cells (line) of our patient (left) and a normal control (right). Quantification by FC was performed by setting the marker against isotype control to determine the percentage of cells staining positive for phospho-STAT5 and its intensity (MCFI). STAT5 phosphorylation in basal conditions is coincident with the isotype control in both the patient and the normal control. B, Total STAT5 protein expression in unstimulated cells (shaded histogram) versus isotype control (line) of our patient (left) and a normal control (right). The percentage of cells positive for STAT5 and MCFI are expressed.

The expression of STAT5 proteins was reduced compared with the normal control (MCFI value: 291 vs 464 ± 62.7 [n = 3]) (Fig 3B).

As a whole, these results suggest that the functional impairment in IL-2-mediated signaling might be the consequence of an alteration in either of the 2 STAT5 proteins.

Molecular Studies

To evaluate the expression of both STAT5a and STAT5b proteins, immunoblot analysis was performed by using an anti-STAT5 mAb and 7% sodium dodecyl sulfate/polyacrylamide gels that discriminate STAT5a (91 kilodaltons) from STAT5b (84 kilodaltons) proteins, as shown in Fig 4A. The STAT5b protein is not detected at all in the patient's sample, whereas the related STAT5a protein level is comparable to those found in the normal B-cell line. The absence of STAT5b was confirmed with an immunoblot analysis using a specific anti-human STAT5b mAb (Fig 4B).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4

Immunoblot analysis of STAT5 expression in our patient's and a normal control's EBV immortalized B-cell lines using anti-human-STAT5-mouse mAb (A) and anti-human-STAT5b-mouse mAb (B) in 7% sodium dodecyl sulfate/polyacrylamide gels. The STAT5b protein is not detected at all in the patient's sample, whereas the related STAT5a protein is detected at levels comparable to those found in the normal B-cell line.

To determine the genetic basis for the apparent lack of immunodetectable STAT5b, STAT5b coding exons were sequenced. The results revealed a novel single-nucleotide change, a homozygous C→T transition, that resulted in a nonsense mutation at codon 152 (R152X). The parents and brother were heterozygous for this mutation. The stop-codon mutation in the STAT5b gene is in exon 5, which encodes the coiled-coil domain, very close to the NH₂-terminal domain, predicting that the protein should not be immunodetectable.

DISCUSSION

The studies described for this patient lend support, for the first time, to the relevance of STAT5b not only in the human GH signaling cascade but also in the cytokine-mediated immune response. We describe here a novel mutation in the STAT5b gene in a patient with a combined phenotype of GH insensitivity and immunodeficiency secondary to lymphocyte functional alterations.

The nonsense mutation, located in exon 5 of the STAT5b gene, predicts a severely truncated protein that retains only the NH₂-terminal domain. Because it lacks the tyrosine-activation domain, it is expected to be devoid of any biological activity. Immunoblot analysis failed to detect the STAT5b protein, probably because the mutant STAT5b product suffers a rapid degradation or it is not expressed at all.

As in the 2 patients described previously,^5,6 basal normal-to-high concentration of serum GH and extremely low serum IGF-I and IGFBP-3 levels were found in this case. Moreover, exogenous GH treatment did not further increase serum IGF-I or IGFBP-3 levels, confirming that STAT5b abnormality affects GHR signaling cascade. Similarly to the other 2 cases with STAT5b gene mutations^5,6 and to those observed in GHR deficiency,⁷ near-normal size at birth and severe postnatal growth failure were observed. Therefore, this postnatal growth pattern reinforces the relevance of the GH-GHR-STAT5b-IGF-I pathway in human postnatal growth. In this regard, the STAT5b protein in STAT5 knock-out models is also essential for GH biological response in terms of postnatal growth and IGF-I expression. However, the postnatal growth observed in the patients is in contrast to that of mice, in which only male animals are affected.⁴

The persistently high hyperprolactinemia suggests a state of prolactin resistance. In this regard, prolactin specifically stimulates phospho-STAT5b in cultured hypothalamic dopaminergic neurons.²⁶ In addition, a specific intracellular pathway of prolactin-negative feedback in the hypothalamus is mediated by STAT5b, as it has been demonstrated in STAT5b knock-out mice, showing grossly elevated serum prolactin concentrations.²⁷ Therefore, the patient's hyperprolactinemia suggests that STAT5b plays a role in mediating the negative feedback action of prolactin also in humans.

The relevant role of STAT5 proteins in the immune system thus far has been defined only in mouse models. Similarly to the 2 patients reported, in our case clinical signs of immunodeficiency were clearly found, because she suffered from recurrent pulmonary infections, persistent herpesvirus infection, chronic eczema, and diarrhea. However, the immune system abnormalities in the 2 patients with STAT5b gene mutations already reported have not been well characterized.

In our patient, similar to the STAT5b knock-out¹⁶ mouse, phenotypic alterations, moderate T-cell lymphopenia, and highly reduced NK and γδ T cells were found. Moreover, the T-cell–hyperactivated status resembles the STAT5a/b double–knock-out mice¹¹ and, interestingly, also the knock-out mice for IL-2Rβ, the signaling-IL-2R chain that serves as a STAT5 docking site.²⁸

T cells were functionally defective, which was clearly documented by the in vitro low proliferative response to mitogens and antigens. Even in the presence of rIL-2, no proliferation improvement was observed. These results reflect a generalized defect in T-cell functions that could be a consequence of both defective STAT5 signals through T-cell receptor (TCR)²⁹ and IL-2R. Moreover, resistance to IL-2 takes place even in the presence of normal amounts of STAT5a and one half of IL-2–STAT5 phosphorylation, which reinforces the concept that STAT5a and STAT5b are not completely redundant.³ Thus, different from mice,¹¹ a full STAT5 function would be necessary in humans for a complete immune response.

On the other hand, we observed that the expression of IL-2Rα chain (CD25) could not be sustained in our patient's T cells in culture even in the presence of rIL-2 (data not shown). The fact that IL-2Rα is controlled by IL-2 through STAT5 proteins^16,30 suggests that the expression of a high-affinity receptor is also affected, and it might have an influence in the lack of response to IL-2.

Nevertheless, the possible impact of the absence of STAT5b on impaired thymic function should also be considered in the context of T-cell lymphopenia, skewed CD45RA/CD45RO ratio, proliferative defects, and chronic infection. Additional studies designed to evaluate thymic output could clarify this point.

Contrary to T cells, B lymphocytes were normal in number and function, suggesting that STAT5b protein is mainly relevant in mediating T-cell–lineage functions. However, the presence of hypergammaglobulinemia and anti-nuclear antibodies, even without clinical autoimmune disease, suggests a deregulation of B cells, probably as a result of an impaired control by defective T cells.

Several evidences show that IL-2 is not only an important growth factor for T cells through TCR stimulation but also exhibits a critical role in the maintenance of in vivo tolerance to self-antigens, because spontaneous accumulation of activated T cells and autoimmunity develop in IL-2³¹ or IL-2Rβ¹⁴ knock-out models or downstream signaling-component knock-out models such us STAT5a/b^−/− mice.^9,11 However, the role of STAT5 proteins in the regulation of the immune response in humans has not been clearly established. One of the mechanisms for maintenance of peripheral tolerance by IL-2 is through the generation of CD4⁺CD25⁺ Tregs.³² In this regard, it has been described in mice that STAT5 protein signaling is required for IL-2 Treg development and homeostasis.^17,18 In knock-out mouse models for STAT5a/b, the reduction of this T-cell subset was considered to contribute to the loss of tolerance detected in a group of them. The reduction in Tregs correlated with a high rate of cell death in this population.⁹

Our findings are compatible with a deregulation of the mechanisms of tolerance in our patient, such as the detection of high titers of antinuclear antibody, hypergammaglobulinemia, and T-cell activation. However, hypergammaglobulinemia and T-cell activation could also be secondary to infections. Moreover, on the basis of the characteristics of the pulmonary disease of the previously reported patients,^5,6 we suspected a similar chronic pulmonary illness as a consequence of lymphocytic infiltration. Because a significant decrease in Treg number that might contribute to this phenotype was found, we propose that in humans complete STAT5 function is required for establishing normal tolerance, although additional studies are necessary to confirm this proposal.

Similarly to STAT5 knock-out mice,^11,16 a severe reduction in our patient's NK-cell number was found, suggesting that either development or homeostasis of these cells in the periphery depend on STAT5b. For both mechanisms, dependence on the IL-15/STAT5 pathway seems to be crucial.³³ However, in contrast to the STAT5b knock-out model, the patient's NK cells showed a response to IL-2, increasing NK cytotoxic function, and upregulating perforin. Although this could be secondary to the activation of an alternative pathway,³⁴ the recurrent herpetic infections in our patient suggest an inadequate NK surveillance, considering the crucial role of these cells in innate immune defense against viruses.³⁵ Moreover, primary immunodeficiencies with selective deficiency in NK cells suffer mainly infections with herpes family viruses.^36,37

CONCLUSIONS

We have demonstrated T-cell–function defects, a low number of NK and γδ T cells, IL-2 signaling defects at the T-cell level but preserved in NK, and deregulation of homeostasis (probably related to a decreased number of Tregs) in a patient with GH insensitivity resulting from a novel mutation in the STAT5b gene.

This mutation results in alterations in the signaling pathways of GHR, cytokine receptors related to T lymphocytes such as IL-2 receptor, and probably prolactin receptor. This is explained biologically by the fact that STAT5b is a signal transduction protein shared by all these factor receptors.