Dopamine β-hydroxylase deficiency impairs cellular immunity

Abstract

Norepinephrine, released from sympathetic neurons, and epinephrine, released from the adrenal medulla, participate in a number of physiological processes including those that facilitate adaptation to stressful conditions. The thymus, spleen, and lymph nodes are richly innervated by the sympathetic nervous system, and catecholamines are thought to modulate the immune response. However, the importance of this modulatory role in vivo remains uncertain. We addressed this question genetically by using mice that lack dopamine β-hydroxylase (dbh^−/− mice). dbh^−/− mice cannot produce norepinephrine or epinephrine, but produce dopamine instead. When housed in specific pathogen-free conditions, dbh^−/− mice had normal numbers of blood leukocytes, and normal T and B cell development and in vitro function. However, when challenged in vivo by infection with the intracellular pathogens Listeria monocytogenes or Mycobacterium tuberculosis, dbh^−/− mice were more susceptible to infection, exhibited extreme thymic involution, and had impaired T cell function, including Th1 cytokine production. When immunized with trinitrophenyl-keyhole limpet hemocyanin, dbh^−/− mice produced less Th1 cytokine-dependent-IgG2a antitrinitrophenyl antibody. These results indicate that physiological catecholamine production is not required for normal development of the immune system, but plays an important role in the modulation of T cell-mediated immunity to infection and immunization.

Norepinephrine (NE) and epinephrine (EP) play important roles in energy balance, thermoregulation, cardiovascular tone, behavior, and the stress response (1–3). These catecholamines also modulate lymphocyte function in vitro and may preferentially impair the generation and function of interferon γ (IFN-γ)-producing Th1 T cells (4–9), which mediate cellular immunity and defense against intracellular pathogens (10, 11). Sympathetic nerve terminals are in close apposition to T lymphocytes and antigen-presenting cells in primary and secondary lymphoid organs, and the concentration of NE in these tissues is much greater than in blood (12, 13). These findings suggest that NE may play a physiological role in immune regulation. However, the conclusions of studies using pharmacologic, surgical, and behavioral approaches to address the effects of NE and stress on the immune response to infection and immunization in vivo have reached varying conclusions (9, 12, 14, 15).

Dopamine β-hydroxylase (DBH) catalyzes the conversion of dopamine to NE (3). Mice in which the dbh gene has been disrupted (dbh^−/− mice) cannot produce NE or EP, but instead produce dopamine as the terminal product of catecholamine biosynthesis. Compared with wild type or dbh^+/− heterozygous controls, dbh^−/− mice have altered metabolism, thermoregulation, cardiovascular tone, and maternal behavior (1, 2, 16). We reported previously that 32% of dbh^−/− mice died during adolescence (16). We subsequently found that these premature deaths did not occur when dbh^−/− mice were housed in a specific pathogen-free (SPF) environment, suggesting that dbh^−/− mice may be more susceptible to infection. To address this possibility, we evaluated their immunological status and their response to infection or immunization. When housed in SPF conditions, and in the absence of an infectious challenge, the immunological status of dbh^−/− mice appeared to be normal. However, when they were infected with Listeria monocytogenes or Mycobacterium tuberculosis, or were immunized with a T cell-dependent antigen, dbh^−/− mice were more susceptible to infection, had impaired T cell function, and had impaired Th1 T cell-dependent-IgG2a antibody production.

MATERIALS AND METHODS

Mice and Facilities.

All mice used in this study were hybrids of the 129/SvCPJ and C57BL/6J inbred strains containing on average 50% of each background. Heterozygous dbh^+/− mice were used as controls because these mice have normal catecholamine levels (2). Mice were rescued during prenatal development by treatment with l-threo-3,4,dihydroxyphenylserine as described (16). Sex-matched dbh^−/− and heterozygous littermate control mice were used between 3 and 9 wk of age for initial in vitro lymphocyte proliferation studies and between 3 and 6 months of age for infection and immunization studies. Mice were bred and housed at the University of Washington under SPF conditions, unless indicated otherwise.

Flow Cytometry.

Analysis of lymphoid populations was accomplished by staining thymus or spleen cell suspensions with flurochrome- or biotin-conjugated mAbs specific for surface proteins found on immature and mature T and B cells, which were prepared in our laboratory or purchased from PharMingen. Biotin-conjugated antibodies were detected with streptavidin-flurochrome conjugates. Stained suspensions were analyzed on a FACScan flow cytometer and analyzed by using cellquest software (Becton-Dickinson).

Experimental Infections.

For L. monocytogenes infections, mice were injected i.p. with 1 × 10⁵ colony-forming units (CFU) in 100 μl of sterile 0.9% saline. The number of viable organisms administered was confirmed by culturing serial dilutions on trypticase soy agar (TSA, Becton-Dickinson). Mice were exposed to an aerosol of 5 × 10⁷ M. tuberculosis (Erdman strain) in 10 ml of sterile PBS/0.05% Nonidet P-40 for 30 min, which deposited ≈400 bacilli in the lungs. At the indicated times, organs were removed, homogenized, serially diluted in PBS/0.05% Nonidet P-40, and cultured on TSA for L. monocytogenes or on 7H10 agar for M. tuberculosis.

Lymphocyte Cultures.

Spleen or lymph node cells were prepared and cultured in serum-free HL-1 medium (BioWhittaker) in 24-well tissue culture dishes as described (17). Cells were stimulated with optimal concentrations of M. tuberculosis culture filtrate proteins (CFP, Corixa, Seattle, WA) or anti-CD3 mAb (145–2C11). Supernatants were harvested after 72 hr, filter-sterilized, and stored at −80°C until assayed for IFN-γ , tumor necrosis factor α (TNF-α), interleukin (IL) 4, and IL-10 by ELISA (Genzyme).

Immunization and Antibody Titers.

Mice were immunized i.p. with 50 μg of trinitrophenyl (TNP)₁₄-keyhole limpet hemocyanin in alum. Sera were collected from mice before and 21 days after immunization. Antibody titers were determined by ELISA, using microtiter plates coated with TNP-BSA and horseradish peroxidase-conjugated antibodies specific to murine IgM, IgG1, and IgG2a (Southern Biotechnology Associates). Endpoint titers were taken as the last dilution at which the OD₄₀₅ exceeded the preimmune value by 0.2 for IgM and 0.1 for IgG1 and IgG2a.

Catecholamine and Hormone Determinations.

After decapitation, blood was put into tubes containing heparin, ascorbic acid, and aprotinin on ice. Plasma was isolated by centrifugation at 4°C, and samples were immediately frozen at −80°C. Catecholamines and corticosterone concentrations were measured by RIA by R. Veith at the Veterans Administration Puget Sound Health System (Seattle, WA) and by J. Licinio at the National Institute of Mental Health (Bethesda, MD), respectively. Prolactin levels were determined by an RIA (AniLytics, Gaithersburg, MD).

RESULTS AND DISCUSSION

Basal Immunological Status of dbh^−/− Mice.

The immune system of dbh^−/− mice housed in SPF conditions appeared to be normal. The total number of leukocytes and the percentages of granulocytes, monocytes, and lymphocytes in the blood (not shown), and the numbers and subpopulations of cells in the thymus (Fig. 1b) and spleen (not shown), were similar to age- and sex-matched dbh^+/− controls. Splenocyte and thymocyte proliferation in response to T cell (Con A and anti-CD3) and B cell (lipopolysaccharide) mitogens, splenocyte production of Th1 (IFN-γ) and Th2 (IL-4) cytokines, and the delayed-type hypersensitivity response to trinitrochlorobenzene were similar in dbh^−/− and control mice (not shown). These results indicated that dbh^−/− mice do not have intrinsic developmental or functional immune defects. In contrast, the thymuses of surviving dbh^−/− mice housed in non-SPF conditions showed marked involution (Fig. 1a). The most marked reduction was observed in the immature CD4⁺CD8⁺ (coreceptor double-positive) thymocytes, which turn over rapidly (every 3–4 days) (18) and are the population most affected in stress- or corticosteroid-induced thymic involution (19). This finding suggested that environmental events (most probably infection) in the non-SPF facility led to involution of the lymphoid organs and increased mortality in dbh^−/− mice. This possibility was explored by infecting or immunizing mice housed in SPF conditions.

Figure 1.

Thymic cellularity in dbh^−/− and dbh^+/− (control) mice. The percentage of CD4⁺, CD8⁺, CD4⁺CD8⁺ (DP), and CD4⁻CD8⁻ (DN) thymocytes was determined by flow cytometry. Results from representative mice are shown as two-dimensional contour plots. Mice were housed under non-SPF conditions (a), were housed under SPF conditions and were not infected (b), or were studied on day 5 of primary (c) or on day 3 of secondary (d) L. monocytogenes infection; the total numbers of cells recovered from these representative mice are shown below the plots. Overall, 4-wk-old dbh^−/− and dbh^+/− mice housed in SPF conditions had similar numbers of thymocytes (1.6 ± 0.3 × 10⁸ vs. 1.7 ± 0.4 × 10⁸ mean ± SEM, n = 6), whereas the numbers of thymocytes in dbh^−/− mice housed in non-SPF conditions were significantly lower than in dbh^+/− mice (3.6 ± 2.0 × 10⁷ vs. 1.6 ± 0.2 × 10⁸, P < 0.005, n = 5–6). The numbers of thymocytes were similar in 3- to 6-month-old uninfected SPF dbh^−/− and dbh^+/− control mice (6.2 ± 1.5 × 10⁷ vs. 5.0 ± 0.6 × 10⁷, n = 5), and in mice during secondary L. monocytogenes infection (5.2 ± 0.7 × 10⁷ vs. 5.6 ± 0.3 × 10⁷, n = 3–5), but were significantly different during primary L. monocytogenes infection (2.5 ± 1.3 × 10⁶ vs. 2.3 ± 0.7 × 10⁷, respectively, P < 0.05, n = 4–5).

Response of dbh^−/− Mice to Infection and Immunization.

After primary infection with L. monocytogenes, the numbers of bacteria were increased in the livers of dbh^−/− mice throughout the course of infection and were higher in their spleens by day 5 (Fig. 2A). Deaths occurred more frequently in dbh^−/− mice between days 5 and 8 (Fig. 2A). The apparent clearance of bacteria in dbh^−/− mice between days 5 and 8 may be, at least in part, an artifact caused by deaths of the dbh^−/− mice with high bacterial burdens during this time. When rechallenged with L. monocytogenes 3 wk after acute infection, the numbers of bacteria were again greater in dbh^−/− mice than in controls (Fig. 2B), though none of the dbh^−/− mice died after secondary challenge.

Figure 2.

dbh^−/− mice have increased numbers of bacteria in their livers and spleens after infection with L. monocytogenes. (A) Mice were infected i.p. with 1 × 10⁵ CFU L. monocytogenes, and the numbers of bacteria in their livers and spleens were determined at the indicated days. Differences between the groups were analyzed by ANOVA. (B) In a separate experiment, mice were infected and euthanized 5 days after primary infection or allowed to control the primary infection and challenged 21 days later with 1 × 10⁵ CFU of L. monocytogenes. Differences between the groups were analyzed by Student’s t test. The tables below each panel show the frequency of deaths for the total number of mice for each day (dead/n). Three dbh^−/− mice died before sacrifice on day 5 and one died before secondary infection (total deaths = 4 of 11). When spleens from mice that died were evaluated (including each of the mice that died in B and a subset of the mice that died in A), high numbers of bacteria (>10⁸ CFU/g) were found, but these were not included in the calculations of organ bacterial densities shown. Values are mean log₁₀ CFU/g tissue ± SEM.

Innate immunity helps to control the growth of L. monocytogenes in the first 3 days after acute infection. Thereafter, antigen-specific, T-cell-mediated immunity eliminates active infection and protects from secondary infection (20, 21). Protection is mediated in part through the production of IFN-γ and TNF-α. Resistance to primary and secondary infection was impaired in dbh^−/− mice, suggesting that they may have defects in the T cell response to infection. Consistent with this finding, the production of cytokines by anti-CD3-stimulated splenocytes from dbh^−/− mice during primary and secondary L. monocytogenes infection was markedly reduced compared with controls (Table 1). In contrast to the depressed cytokine production by dbh^−/− splenocytes, the numbers of splenocytes, the fraction of splenocytes that were CD4⁺ and CD8⁺ T cells, and splenocyte proliferation were similar (Table 1). In sham-infected dbh^−/− mice, there was little (primary) or no (secondary) reduction in cytokine production, and the numbers and proliferation of splenocytes from dbh^−/− and control mice were similar (Table 1). Thymic involution was observed in dbh^−/− mice during primary L. monocytogenes infection (Fig. 1c) and resembled that seen in dbh^−/− mice housed in non-SPF conditions (Fig. 1a). Whether the exaggerated thymic involution contributed to the poor outcome of acute L. monocytogenes infection in dbh^−/− mice is unclear, but this involution did not appear to contribute to their inability to control secondary infection, because thymic cellularity of dbh^−/− and control mice was similar at that time (Fig. 1d). Neutrophil recruitment, which helps control infection caused by L. monocytogenes in the liver but not in the spleen during the first 24 hr of primary infection (21), appeared to be reduced in the livers of dbh^−/− mice (not shown). Reduced neutrophil recruitment may have contributed to the higher numbers of bacteria in their livers at 24 hr (Fig. 2A). Nonetheless, the predominant and persistent defect in dbh^−/− mice appeared to be the marked impairment in cytokine production by T cells after infection.

Table 1.

Cytokine production and proliferation by splenocytes from mice infected with L. monocytogenes

	dbh Genotype	IFN-γ, pg/ml	TNF-α, pg/ml	IL-10, pg/ml	Proliferation, cpm	Spleen cell no., CD4/CD8
Primary infection	+/−	16,200*	1,882	3,770	10,662† (5.0)	3.5  ×  107 (1.9)‡
	−/−	4,918	334	180	9612 (3.4)	2.3  ×  107 (1.4)
Primary sham	+/−	2,061	2,192	3,680	17,795 (17.7)	1.8  ×  107 (2.2)
	−/−	1,238	599	1,120	17,223 (27.7)	1.6  ×  107 (2.0)
Secondary infection	+/−	12,083	316	650	20,344 (17.1)	1.5  ×  107 (2.0)
	−/−	525	<30	100	12,109 (25.3)	1.0  ×  107 (2.8)
Secondary sham	+/−	2,730	466	620	25,501 (27.6)	1.4  ×  107 (1.9)
	−/−	3,207	748	564	25,317 (28.3)	1.2  ×  107 (1.9)

Mice were inoculated i.p. with 1 × 10⁵ CFU of L. monocytogenes or were given an identical volume of saline i.p. (sham). Cells from the dbh^−/−(n = 3–4) and dbh^+/− (n = 5–6) mice shown in Fig. 2B were pooled according to group and 2.5 × 10⁶ were incubated in microtiter wells in medium alone or with anti-CD3. 

^*Cytokines values are the mean of triplicate wells assayed in duplicate. Unstimulated values and all values for IL-4 (not shown) were below the limits of detectability. 

^†Mean ³H-thymidine uptake of triplicate wells (stimulation index, anti-CD3 cpm/unstimulated cpm). 

^‡Mean number of cells recovered from spleens (ratio of CD4⁺ to CD8⁺ T cells). 

Further support for this notion, and for a more selective impairment of Th1 cytokine production, was provided by studies in dbh ^−/− mice that were infected by aerosol with M. tuberculosis (Table 2). Like L. monocytogenes, protection against M. tuberculosis depends on Th1 T cells that produce IFN-γ and TNF-α (22, 23), production of which is inhibited by the Th2 cytokine IL-10 (10, 11). When evaluated 4 wk after infection, splenocytes from dbh^−/− mice produced less IFN-γ and TNF-α and more IL-10 than splenocytes from controls in response to stimulation with M. tuberculosis CFP and anti-CD3 (Table 2). Results with cells from draining lymph nodes were similar, with the exception that IFN-γ and TNF-α production by dbh^−/− mice was reduced only in response to CFP. Although this experiment was designed to assess the T cell response rather than microbiological outcome, the numbers of M. tuberculosis bacilli in the lungs of dbh^−/− mice were somewhat greater than in the controls (8.0 ± 0.3 vs. 7.4 ± 0.2 log₁₀ CFU/lung, n = 8 per group).

Table 2.

Cytokine production by cells from mice infected with M. tuberculosis

Cell source	Stimulus	dbh Genotype	IFN-γ, pg/ml	TNF-α, pg/ml	IL-10, pg/ml
Spleen	Anti-CD3	+/−	3,358*	844	<30
		−/−	851	<30	884
	CFP	+/−	1,020	<30	<30
		−/−	297	<30	<30
Lymph nodes	Anti-CD3	+/−	3,938	931	<30
		−/−	4,552	1,379	700
	CFP	+/−	1,009	781	<30
		−/−	432	283	<30

Mice (eight per group) were infected with M. tuberculosis by aerosol delivering ∼400 CFU to the lungs. Four weeks later, cells from mice in each group were pooled and 2.5 × 10⁶ were incubated in microtiter wells in medium alone or with anti-CD3 or anti-CD3 or mycobacterial CFP. 

^*All values are the mean of triplicate wells assayed in duplicate. Unstimulated values and all values for IL-4 (not shown) were below the limits of detectability. 

To determine whether the response of dbh^−/− mice to immunization also differed from controls, mice were immunized with TNP-keyhole limpet hemocyanin, the antibody response to which is T cell dependent. Compared with dbh^+/− controls, dbh^−/− mice had lower titers of IFN-γ-dependent IgG2a (10) anti-TNP antibody (2.9 ± 0.2 vs. 3.8 ± 0.2 log₁₀ titer, P = 0.01), but similar amounts of IgG1 (5.0 ± 0.2 vs. 5.3 ± 0.2 log₁₀ titer) and IgM (not shown) anti-TNP antibody. Thus, Th1 responses were impaired in dbh^−/− mice in response both to immunization with TNP-keyhole limpet hemocyanin and infection with M. tuberculosis.

The Hypothalamic-Pituitary-Adrenal (HPA) Axis in dbh^−/− Mice.

The sympathetic nervous system and HPA axis influence each other’s actions and work in concert to regulate the host response to a variety of stresses, including infection (9, 12, 15). The HPA axis responds to stress by inducing the production of glucocorticoids, which inhibit multiple T cell functions, preferentially impairing Th1 T cell responses (15, 24). To address the role of the HPA axis in the impaired immunity of dbh^−/− mice, we measured morning corticosterone concentrations. These concentrations were somewhat higher (P = 0.08) in uninfected dbh^−/− mice, and significantly higher (P = 0.02) in dbh^−/− mice during primary, but not secondary, L. monocytogenes infection (Table 3). Thus, excess corticosterone production may have contributed to the impaired response to primary infection with L. monocytogenes in dbh^−/− mice. However, dbh^−/− mice had a markedly impaired response to secondary infection with L. monocytogenes (Fig. 2B), even though corticosterone concentrations were not elevated at this time. This finding suggests that impaired cellular immunity in dbh^−/− mice was not principally caused by increased corticosterone production, although it is possible that the development of protective memory T cells was impaired in part by elevated corticosterone concentrations during the primary infection.

Table 3.

Catecholamine, prolactin, and corticosterone concentrations in L. monocytogenes- or sham-infected mice

	Genotype	Numbers	NE	EP	Dopamine	Prolactin	Corticosterone
Primary infection	+/−	5	6.0  ±  0.6	10.8  ±  1.6	0.7  ±  0.1	4.0  ±  0.4	86.3  ±  19.7
	−/−	4	<0.1	0.7  ±  0.2	9.7  ±  3.6	2.9  ±  0.5	366.7  ±  101.2†
Primary sham	+/−	5	5.3  ±  0.5	7.9  ±  1.1	0.9  ±  0.2	5.0  ±  1.0	49.2  ±  18.0
	−/−	5	<0.1	1.0  ±  0.2	12.0  ±  1.7	7.1  ±  1.0	114.2  ±  26.9‡
Secondary infection	+/−	5	2.9 ± 0.3	4.8  ±  0.5	0.4  ±  0.1	nd	54.4  ±  13.1
	−/−	3	<0.1	0.3  ±  0.2	4.5  ±  2.0	nd	68.4  ±  57.8
Secondary sham	+/−	5	3.7 ± 0.4	4.4  ±  0.7	0.8  ±  0.1	nd	116.9  ±  23.9
	−/−	5	<0.1	0.6  ±  0.2	7.6  ±  2.0	nd	77.0  ±  28.4

Mice were infected i.p. with L. monocytogenes or sham-infected (i.p. saline) as described in Fig. 2. Plasma was collected day 5 postprimary and day 3 postsecondary infection. Values are the mean ± SEM (ng/ml). We have not previously detected EP in dbh^−/− mice using HPLC and electrochemical detection (2), suggesting that the current values for EP in dbh^−/− mice reflect cross-reactivity of dopamine in the RIA used in the current studies. nd, not done. 

^†P = 0.02,  

^‡P = 0.08, by two-tailed Student’s t test. Other comparisons for corticosterone and prolactin concentrations, P > 0.1. All catecholamine values differed significantly between dbh^−/− and dbh^+/− mice (P < 0.01). 

Dbh^−/− mice have essentially no circulating NE or EP, but dopamine levels are increased ≈10-fold (Table 3) (2). Dopamine can inhibit prolactin production, thereby blocking the enhancing effect of prolactin on T cell function (25), including protection against infection with L. monocytogenes (26). However, prolactin concentrations were similar in control and dbh^−/− mice both before and after infection (Table 3), arguing against dopamine-mediated inhibition of prolaction secretion as a mechanism for impaired cellular immunity.

The current studies provide genetic and biochemical evidence that DBH deficiency has a profound influence on the efficacy of the immune response and, in particular, Th1 T cell-dependent responses. Despite the absence of NE and EP, and increased dopamine and basal corticosterone concentrations, the development and function of the immune system in dbh^−/− mice was normal in the absence of an infectious or immunological challenge in vivo, but markedly abnormal after challenge. This finding supports the notion that a properly regulated response involving the sympathetic nervous system and the HPA axis is important for immunological as well as behavioral, cardiovascular, and metabolic adaptation to stress (12, 14, 15). Given the high local concentrations of NE and the close apposition of sympathetic nerve terminals to T cells in lymphoid organs (12), it is likely that locally released rather than circulating NE plays the major role in immune modulation. Aberrant catecholamine biosynthesis also affected production of corticosterone by the HPA axis, which may have contributed to the defective immune response in dbh^−/− mice. Whether the latter effect was caused by chronic NE deficiency, excess dopamine, or both will require further evaluation. Recent data suggest that leptin-deficient mice have host defense defects similar to those in dbh^−/− mice (27). Leptin deficiency reduces sympathetic outflow (28), which may contribute to the immunological impairment in leptin-deficient mice. However, dbh^−/− mice have normal leptin concentrations (1), arguing against leptin deficiency as a mechanism for their impaired immunity. The current results indicate an intimate relationship between the immune and neuroendocrine systems and support the notion that NE plays an important role in the regulation of host defenses.

ABBREVIATIONS

CFP

culture filtrate proteins

CFU

colony-forming units

DBH

dopamine β-hydroxylase

EP

epinephrine

HPA

hypothalamic-pituitary-adrenal

IFN-γ

interferon γ

IL

interleukin

NE

norepinephrine

SPF

specific pathogen-free

TNP

trinitrophenyl

TNF-α

tumor necrosis factor α