Undernutrition is associated with perturbations in T cell-, B cell-, monocyte- and dendritic cell- subsets in latent Mycobacterium tuberculosis infection

Undernutrition, as described by low body mass index (BMI), is a foremost risk factor for the progression of active Tuberculosis (TB). Undernutrition is also known to impact the baseline frequencies of innate and adaptive immune cells in animal models. To verify whether undernutrition has any influence on the baseline frequencies of immune cells in latent Mycobacterium tuberculosis infection (LTBI), we examined the frequencies of T cell-, B cell, monocyte- and dendritic cell (DC)- subsets in individuals with LTBI and low BMI (LBMI) and contrasted them with LTBI and normal BMI (NBMI) groups. LBMI was characterized by decreased frequencies and absolute cell counts of T cells, B cells and NK cells in comparison with NBMI. LBMI individuals demonstrated significantly enhanced frequencies of naïve and effector CD4+ and CD8+ T cells and significantly decreased frequencies of central memory, effector memory CD4+ and CD8+ T cells and regulatory T cells. Among B cell subsets, LBMI individuals demonstrated significantly diminished frequencies of naïve, immature, classical memory, activated memory, atypical memory and plasma cells. In addition, LBMI individuals showed significantly decreased frequencies of classical monocytes, myeloid DCs and plasmacytoid DCs and significantly increased frequencies of intermediate and non-classical monocytes and myeloid derived suppressor cells. BMI exhibited a positive correlation with B cell and NK cell counts. Our data, therefore, demonstrates that coexistent undernutrition in LTBI is characterized by the occurrence of a significant modulation in the frequency of innate and adaptive immune cell subsets.


Introduction
Globally, Tuberculosis (TB) continues as the foremost reason for infection related illness and death. In 2017, the World Health Organization reported 10 Globally, approximately 23% of the population are with LTBI [1]. Among individuals with LTBI, only about 5 to 10% develop active TB during their life time and this conversion occurs due to breakdown in the protective immune mechanism [2]. Both nutrition and immunity are strongly interlinked. Innate and the adaptive immune systems are influenced by nutritional status and these immune cells have a role in nutritional immunology. Phagocytosis, T cell numbers and cell-proliferation response to mitogens are affected due to undernutrition [3,4]. Nutritionally compromised individuals, who had vaccination, also exhibited diminished specific antibody titers [5]. Undernutrition dampens the cellmediated immunity and predisposes individuals to become more vulnerable to active TB disease [4,6,7]. Several developing countries have high TB burdens concomitant with undernutrition. Undernutrition has the highest population attributable fraction (27%) of any risk factor in many countries with the highest TB burden [8][9][10].
Nonetheless, in humans, the functions of innate and adaptive immune cells in undernourished individuals with LTBI have not been explored in detail. Very few studies have shown data on the immunological mechanism of predisposition from latent to active TB disease. We postulate that undernutrition could diminish the cellular responses in LTBI and thus weaken the immune system and which in turn cause individuals with LTBI to be more prone to active TB disease. To study the effect of undernutrition on LTBI, we compared the frequencies of T cell-, B cell-, monocyte-and dendritic cell (DC)-subsets between LTBI with low BMI (LBMI) group and LTBI with normal BMI (NBMI) group.

Ethics statement
The study protocol was approved by Institutional Review Board of the National Institute of Research in Tuberculosis, Chennai, India (approval no. NCT00375583 and NCT00001230) and as part of the natural history protocol, informed written consent was taken from all study participants.

Study population
We enrolled 60 study participants with LTBI, with 30 participants with LBMI and 30 participants with NBMI between 2015 and 2018 ( Table 1). All the participants were residents of rural villages of Kanchipuram District, Tamil Nadu, South India with an age range from 18 to 65 years. These study participants were all enrolled from a rural population by screening of individuals for BMI and LTBI. We screened a total of 200 participants with LTBI to recruit 30 with LBMI and 30 matched controls (for age and sex) with NBMI. The flow chart is given in S1 Fig. instructions (Qiagen). Anthropometric measurements, consisting of height, weight, and waist circumference (all measured in triplicate), and bio-chemical parameters, including plasma glucose, serum albumin, urea, creatinine, alanine aminotransferase (ALT), aspartate aminotransferase (AST), and HbA1c levels were measured using normalised methods. Low and normal BMIs were described based on the American Heart Association/American College of Cardiology guidelines (LBMI, <18.5 kg/m2; and NBMI, between 18.5 and 24.9 kg/m2). Additionally, undernutrition was confirmed using low serum albumin (<3.4 g/dl) in all the low-BMI individuals. Low BMI individuals with normal albumin were excluded. All participants who were recruited for the study were negative for diabetes, HIV and parasitic infections. Diabetes was defined as an glycated haemoglobin (HbA1c) reading of 6.5% or greater and a random blood glucose of >200 mg/dl, according to the American Diabetes Association criteria. Parasitic infections were ruled out by stool microscopy and serology. This study population was the same as the population used for our previous study on chemokine analysis [12].

Statistical analysis
Based on preliminary analysis, a sample size of 30 participants per arm was determined to be sufficient to detect a cell count difference of 26 with 90% power at the 5% level of significance assuming a pooled standard deviation of 5 to detect differences in the absolute counts of T, B and NK cells. Central tendency was measured using Geometric means (GM). Statistically significant differences between two groups were analysed using non-parametric Mann-Whitney U-test. Multiple comparisons for each set of experiments (T cell subsets; B cell subsets; monocyte subsets and dendritic cell subsets) were individually corrected using the Holm's correction. Thus, p values obtained by the Mann-Whitney U test for each cell subset were corrected by recalculating them using the Holm's correction method. Data analyses were performed using GRAPHPAD PRISM Version 6 (GraphPad, San Diego, CA). The Spearman rank correlation matrix analysis was done using JMP 13 (SAS) software.

Study population characteristics
Based on BMI and serum albumin, LTBI positive individuals were categorized into two groups: (i) LBMI with BMI of <18.5 kg/m 2 and serum albumin levels of <3.4 g/dl and (ii) NBMI with BMI between 18.5 and 24.9 kg/m 2 and serum albumin levels between 3.7 to 4.8 g/ dl. There was no statistically significant difference between the two groups in the levels of random blood glucose, HbA1c, urea, creatinine, ALT and AST levels. The haematological parameters are shown in Table 2. The haematological parameters did not show any significant difference between the two groups. Age or gender did not differ significantly between two groups.

Low BMI is associated with diminished frequencies and absolute counts of T cells, B cells and NK cells
To study the effect of BMI on leukocyte subsets in LTBI, we determined the frequencies and cell count of T cells, B cells and NK cells in LBMI and NBMI individuals with coexistent LTBI. As can be seen in Fig 1A,

LBMI is associated with altered frequencies of CD4 + T cells and CD8 + T cells
Next, to determine the influence of BMI on CD4 + T cell and CD8 + T cell subsets in LTBI, we measured the frequencies of different CD4 + T and CD8 + T cell subsets (naive, central memory,  memory CD8+ T cells (GM of 27.72% in LBMI versus 31.34% in NBMI, p = 0.0398) were significantly diminished when compared with NBMI group. Therefore, LBMI with LTBI is associated by changes in the CD4 + and CD8 + T cell subset frequencies distribution.  Fig 3A depicts an illustrative flow cytometry dot plot of B cell subsets gating strategy. As can be seen in Fig 3B,

LBMI is associated with altered frequencies of monocyte and dendritic cell subsets
Next, to determine the influence of BMI on monocyte and dendritic cell phenotype in LTBI, we first measured the frequencies of monocyte subsets (classical monocytes, intermediate monocytes and non-classical monocytes) in LBMI and NBMI groups. Fig 4A depicts

Relationship between immune cells and BMI
Next, we determined the association between T cell-, B cell-, monocyte-and dendritic cell (DC)-subset frequencies and BMI. As shown in Fig 5,

Discussion
Even though nutrition is documented as being an important element of immunity, the mechanism by which undernutrition affects the immune functions remains unclear [13,14]. Both innate and adaptive immune systems are influenced by the nutritional status. Animal studies revealed that diminished T cell numbers were observed in severe undernutrition or starvation of animal models, which also affect T-cell survival and proliferation [15,16]. Both CD4 + and CD8 + T cell counts were diminished in undernourished humans when compared with normal weight controls [17]. A previous study showed that undernourished individuals exhibited diminished CD4/CD8 ratio and increased numbers of CD4 and CD8 double negative T cells [18]. Schlesinger et al. reported that undernourished mice had reduced NK cell numbers and NK cell activity [19]. In corroboration with this, our data also reveal diminished levels of CD4 + , CD8 + T cell, B cell and NK cell frequencies as well as counts in LBMI individuals with LTBI. Frequencies and counts of T cells, B cell and NK cells were significantly positively correlated with BMI. Thus, low BMI induces alteration in cell-mediated immunity which could potentially affect individuals host response to infection. In humans, the association between undernutrition and homeostatic memory cell maintenance is unknown. Memory cells are critical for the defence from reinfection. The defence mechanism depends on antigen-independent persistence of memory T and B cells primed to respond rapidly upon re-challenge. Memory homeostasis process involves cell survival which relies on nutrient availability [20,21]. There is a paucity of information on how nutrition influences memory T cell homeostasis. Proliferation of memory cells requires dietary protein which might be crucial for maintaining memory T cells. Studies on protein energy malnutrition (PEM) mice models showed that lymphocytic choriomeningitis virus (LCMV)-specific CD8 + T EM were diminished and T EM were not sustained and in tuberculosis model exhibited reduced lung CD4 + T cell responses to mycobacterial antigens [22,23]. It has also been shown that in undernourished animals, nutritional replenishment resulted in reconstitution of memory T cell responses which improves pathogen clearance [22,23]. A study on undernourished mice revealed that there was a significant and swift reduction in the number of thymocytes, splenocytes, and lymphocytes in these mice [24]. More recent data on undernourished mice demonstrated that mice exhibited significant decreased numbers of effector and regulatory T cells and also alteration in the balance of T cell subsets [25,26]. Our data also reveal that significantly diminished frequencies of classical memory, effector memory and regulatory T cell subsets are characteristics of LBMI with LTBI. BMI exhibited negative correlation with naive and effector T cells, while central memory and effector memory T cells exhibited significant positive correlation with BMI. Therefore, our data reveal that low BMI is associated with altered frequencies of memory T cell subsets in LTBI, perhaps enhancing the possibility of progression to active TB disease.
While there is an increasing indication of the importance B cell responses to intracellular pathogens, B cell subsets in TB have not been studied in detail [27]. Studies have shown that both naive and memory B cells are present in tuberculosis granulomas [27]. M.tb-infected lung contains B cells, which may be involved in the process of M.tb antigen presentation to T cells and secrete cytokines and M.tb-specific antibodies [28]. In our study, LBMI group showed significantly decreased percentages of B cell subsets and also the frequencies of B cells gating strategy of plasmacytoid (pDC) and myeloid DCs (mDC). Plasmacytoid DC were defined by the expression of (Lin − HLA-DR + CD123 + ) and myeloid DCs as (Lin − HLA-DR + CD11c + ). ( [29,30]. Abe et al reported that undernourished mice exhibited decreased levels of B lymphocytes and these B cell subsets have impairment in the secretion of immunoglobulin, cytokine and proliferation [31][32][33]. Our data on diminished frequencies of classical memory B cells, activated memory B cells and plasma cells suggests that the capacity of B cells to secrete antibodies in the face of ongoing TB infection might be compromised in these individuals. Our data also suggest that undernutrition impairs the capacity of B cell production, maintenance or proliferation since, all the subsets of B cells are decreased in frequency. Monocytes are the major players in the immune response against TB. Based on the relative expression of CD14 and CD16, the monocyte subsets are divided into classical, intermediate and non-classical monocytes [34]. Classical monocytes consist of about 80-95% of circulating monocytes, known to be main scavenger cells and are highly phagocytic. Intermediate monocytes encompass up to 5% of total monocytes. These subsets plays a role in production of reactive oxygen species (ROS), antigen presentation and are also involved in T cell proliferation and stimulation and inflammatory responses [35,36]. Non-classical monocytes consist of 5-10% of monocytes. These subsets are known as patrolling monocytes and are pro-inflammatory in behaviour and produce inflammatory cytokines in response to infection [36,37]. Our findings reveal that intermediate and non-classical monocyte percentage was significantly decreased in LBMI compared to NBMI group. Moreover, BMI exhibited negative correlation with frequencies of monocyte subsets. The innate immune response functions such as phagocytosis and secretion of reactive oxygen and nitrogen intermediates by macrophages are affected due to undernutrition [31]. Recent studies have reported that M.tb infection alters the CD16 intermediate monocyte subset, implying that this subset supports the intracellular persistence of M.tb [38]. The altered frequencies of monocyte subsets suggest that LTBI with LBMI group are associated with defect in their monocyte functions.
DCs plays the major role in bridging innate and adaptive immune response via their function in encapsulating, processing and presenting antigens. DCs have been categorized as HLADR + cells and based on expression classified into myeloid DCs (CD11c + ) or plasmacytoid DCs (CD123 + ) [39]. Earlier studies stated that migration of DC to the draining lymph node is necessary for the triggering naive T cells in TB infection [40]. Myeloid DCs and plasmacytoid DCs plays key role in stimulation of CD4 + and CD8 + T cells and support the antibacterial activity [41,42]. M.tb-infected mice showed worsening of disease and enhanced survival of M. tb due to excessive deposition of myeloid-derived suppressor cells (MDSC) [43]. Several studies have stated the negative role of MDSC in anti-TB immunity, T cell activation, proliferation, trafficking, regulatory T cell stimulation and T cell cytokine responses [44][45][46]. Our data reveals that the percentage of plasmacytoid and myeloid DC are significantly diminished and in contrast, MDSC frequencies were significantly higher in LBMI in comparison with NBMI group. The functional effects of this diminished phenotypic distribution of DC subsets in LTBI-LBMI individuals requires additional studies. Previous studies have shown that diminished frequencies of DC subsets could lead to impaired production of IL-12, fewer IFNγ T cells and lower T cells proliferation due to undernutrition [31,47]. Our data therefore indicate that low BMI has profound effects on the DC subsets in LTBI.
Our data offers an understanding into the effect of nutrition on the pathogenesis of TB. To our knowledge, this is the first study to examine the impact of low BMI on innate and adaptive correlation rank matrices with r values being indicated by horizontal bars. p and r values were determined by Spearman rank correlation test at 95% confidence intervals by JMP software.
https://doi.org/10.1371/journal.pone.0225611.g005 immune cells subsets in the context of LTBI. Although, our data do not provide any mechanistic conclusions, our data offers an important preliminary substantiation for a major relationship between undernutrition and immune cell composition in a chronic infection. Our study suffers from the limitation of not being a longitudinal study, of not measuring micronutrient levels or body fat composition and of having a modest sample size. Also, the cellular analysis is non-specific and these are not mycobacterial-antigen stimulated cells. Longitudinal studies of low BMI individuals with LTBI should help elucidate the mechanism by which undernutrition promotes the progression from latent to active infection. Nevertheless, our study clearly implies that perturbations in immune cell frequencies and numbers might be involved in the process. In conclusion, by studying the innate and adaptive cellular responses at homeostasis in undernourished individuals with LTBI, our study offers the direct relationship of compromised immune cellular response in individuals with low BMI and LTBI.