Reduced Neuroinflammation and Pain with a Functional Sourdough Bread Enriched with Legumes and Ancient Cereals in a Mouse Model of LPS-Induced Inflammation

Abstract

Nutritional strategies based on sourdough fermented breads with wholemeal ancient grains and legumes are emerging as promising modulators of (neuro)immune processes. This study investigated whether prolonged consumption of a sourdough bread enriched with a mixture of ancient cereals and legumes, commercially available in Italy (Primus^® bread, P^®B), modulates neuroimmune systemic responses to repeated lipopolysaccharide (LPS) challenge in mice. For this study, male C57BL/6J mice were fed for 14 days with either a standard diet (SD) or P^®B. Animals then received intraperitoneal LPS (3 mg/kg/day for 3 days) or vehicle. Body weight and food intake were monitored throughout. Pain-like behaviours were assessed by von Frey, plantar and tail flick tests, and plasma cytokine (32-plex panel), splenocyte and peritoneal macrophage cytokine expression, and expression of pro-inflammatory cytokines in sciatic nerves, dorsal root ganglia (DRG) and the spinal cord were analyzed by Reverse Transcription–quantitative Polymerase Chain Reaction (RT-qPCR). P^®B prevented LPS-induced body weight loss and reduced splenomegaly. Unlike SD mice, which exhibited widespread plasmatic cytokine upregulation, P^®B-fed mice displayed only limited increases Interleukin (IL)-1β, IL-12p40 and Tumor Necrosis Factor (TNF)α. Ex vivo cultures of splenocytes and macrophages confirmed attenuated cytokine overexpression. LPS-induced hypersensitivity to mechanical, thermal and nociceptive stimuli was significantly reduced in P^®B mice. Molecular analyses revealed that the P^®B diet blunted the pro-inflammatory cytokine expression present after LPS challenge in the sciatic nerves and DRG, with partial attenuation in the spinal cord. Our findings highlight the great potential of functional foods as affordable dietary strategies to mitigate systemic immune and neuroimmune dysregulation.

1. Introduction

Nutritional sciences are revealing that what we eat does much more than just provide metabolites: the aliments termed “functional foods” can potently modulate immune responses, oxidative stress, health and permeability of the gut barrier via their influence on its microbiota and ultimately affect neuroimmune communication [1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22]. Diets rich in whole grains, legumes and fibre, particularly those including fermentation of ancient grains, are gaining attention for their ability to provide bioactive compounds, phenolics, resistant starches and diverse protein profiles, which may buffer inflammatory responses and reduce the risk of immune- and neuroimmune-driven pathologies [1,20,21,23,24,25,26,27]. Growing experimental evidence shows that bread or grain/legume-based dietary interventions can positively alter immune status [2,3,8,14,22,28,29,30,31,32,33,34,35,36,37,38,39,40]. For instance, replacing 24% of wheat flour with red lentil flour produced a higher-fibre bread, with balanced amino acid composition, greater polyphenol content and improved antioxidant capacity to influence serum cytokines and intestinal immune cell populations compared to standard wheat bread in aged mice [30]. Recent murine feeding trials with sourdough and high-fibre breads have also demonstrated amelioration of gut microbiome composition, with reduced systemic inflammatory markers compared to white bread, improved barrier function and better metabolic parameters [19]. Human randomized trials corroborate these findings: substituting whole grains for refined grains in overweight or obese subjects for several weeks significantly reduced IL-6 and other pro-inflammatory markers [23]. Moreover, a systematic review of Randomized Controlled Trials (RTCs) reports that whole-grain consumption is associated with reductions in inflammatory markers (e.g., C Reactive Protein, IL-6 and TNFα) in overweight or metabolically challenged individuals [18]. Many observational studies further support this association [1,35]. However, the effects of complex functional breads combining ancient cereals, legume proteins and sourdough fermentation on systemic inflammation and neuroimmune outcomes have never been investigated before in a murine model. We aim to address this gap in the present work by investigating the effects of a bakery product available on the Italian market, commercialized as “Primus^® bread” (P^®B). The manufacturer developed this product using an iterative, heuristic approach, particularly focused on customer feedback (see the Methods Section). The aim was to produce a whole-cereal, protein-enriched and highly nutritious (“healthy”) bread that offered high palatability, the ability to effectively satisfy appetite, and positive, hedonic affective response during and after intake [41,42,43,44]. The final product is composed of flours from organically grown ancient grains and cereals supplemented with legume proteins, all processed via a proprietary sourdough fermentation protocol [45,46,47] (list of ingredients in Methods). This method was selected over faster yeast leavening because of its documented benefits in improving flavour, texture, and digestibility [31,32,33,45,46,48,49,50,51]. The manufacturer’s promotional efforts did not claim or suggest any functional value of the product. The present study was prompted by recurrent unsolicited customer feedback indicating perceived symptomatic relief in inflammation-related conditions after dietary incorporation of P^®B. Recognizing that these reports were anecdotal and susceptible to confounding factors such as random variation or self-induced placebo effects [52], the manufacturer sought to scientifically validate the observed effects. The analysis of the existing literature on the functional effects of sourdough breads, especially the ones combining wholemeal cereals and pulses [1,13,14,17,18,19,20,21,24,25,26,27,30,35,36,46,51,53,54,55,56,57,58,59,60,61,62,63,64] suggested a compelling rationale for further investigation. On these grounds, and drawing on the expertise of our groups, we sought to determine whether a diet based exclusively on P^®B could modulate the systemic inflammation and neuroinflammation induced by injection of bacterial lipopolysaccharide (LPS) in mice. When immune-activating signals such as LPS enter circulation, they trigger a cascade of pro-inflammatory cytokines that can breach peripheral checkpoints, activate glial cells, disrupt the blood–brain barrier and drive neural circuits toward heightened pain sensitivity [65,66]. This experimental model was selected because it reproduces several features common to inflammatory diseases and is widely used to evaluate the efficacy of anti-inflammatory compounds. The parameters we evaluated were the effects on body weight and food intake (tolerability), systemic cytokines, immune cell activation, pain-related behaviours and pro-inflammatory gene expression in peripheral and central nervous tissues [65,66,67,68,69,70,71,72,73]. By combining behavioural, immunological and molecular data, our aim was to determine whether such dietary intervention could significantly interfere with the onset of peripheral inflammation and of pro-inflammatory cytokine overexpression in the peripheral and central nervous systems and with increased pain sensitivity.

2. Results

2.1. Body Weight and Food Intake

Upon arrival at the animal facility, all mice were acclimatized for 7 days and fed the pelleted standard diet (SD). On day 0, half of the animals were kept on the SD (n = 14, Control (CTR) + SD), while the other animals started a P^®B-based diet (n = 14, CTR + P^®B), both ad libitum. We observed that, throughout the experimental protocol, body weight growth and daily amount of food consumption were comparable between CTR + P^®B and CTR + SD mice (Figure 1A,C). This indicated that exclusive consumption of P^®B successfully supported mice’s normal growth pattern. On day 14, half of the animals of both dietary groups were i.p. treated for 3 consecutive days with 3 mg/kg LPS, a dose able to induce significant immune system activation [74].

In LPS-treated mice, irrespective of diet (i.e., in both LPS + SD and LPS + P^®B), we observed slight but significant and progressive body weight loss (Figure 1A,B; CTR + SD vs. LPS + SD, p < 0.001; CTR + P^®B vs. LPS + P^®B, p < 0.01). Nevertheless, body weight loss was significantly higher in LPS +SD mice than in LPS + P^®B (Figure 1B, p < 0.001). In addition, LPS injection also led to a decrease in food consumption; however, this was significant only in LPS + SD mice (Figure 1C; CTR + SD vs. LPS + SD, p < 0.05).

2.2. Plasma Cytokine Levels

No statistically significant changes in plasma cytokine levels were detected in CTR + SD vs. CTR + P^®B animals. Therefore, CTR mice are shown as the average values of all controls. In Figure 2 it is shown that the LPS administration in mice fed with the SD (LPS + SD) led to a significant increase in most plasma pro-inflammatory cytokines (Figure 2A–G; CTR vs. LPS + SD, p < 0.01).

In contrast, LPS-induced inflammation was significantly attenuated in mice fed with a P^®B-based diet. Indeed, in LPS + P^®B mice, only an upregulation of the plasma pro-inflammatory cytokines IL-1β, IL-12p40 and TNFα (Figure 2A,D,G) was observed (CTR vs. LPS+ P^®B, p < 0.001); however, these values remain significantly lower than those observed in LPS + SD mice (p < 0.01). In addition, plasma levels of the other pro-inflammatory cytokines analyzed were not upregulated compared to CTRs (CTR vs. LPS + P^®B, p > 0.05). Furthermore, in both LPS groups (LPS + SD and LPS + P^®B), there was a comparable upregulation of the anti-inflammatory cytokines IL-10, IL-4 and IL-13 compared to CTRs (Figure 2H–J). The levels of the other analytes investigated are shown in Appendix A (Figure A1). A general upregulation of pro-inflammatory cytokines (Figure A1, panels A–F), growth factors (Figure A1, panels G–K), enzymes (Figure A1, panel L) and chemokines (Figure A1, panels M–U) was observable in LPS mice regardless of the diet.

2.3. Cytokine Expression Levels in Splenocyte and Macrophage Cultures

To further evaluate possible diet-related differences in LPS-induced systemic inflammation, spleen weight changes were measured (Figure 3A). In LPS + SD mice, spleen weight increased 0.94 ± 0.20-fold (CTR vs. LPS + SD, p < 0.001), while for LPS + P^®B mice the increase was 0.45 ± 0.16-fold (CTR vs. LPS + P^®B, p < 0.01). Although a significant increase in spleen weight was observed in both LPS groups, splenomegaly was more severe in LPS + SD mice compared to those with the P^®B diet (LPS + SD vs. LPS + P^®B, p <0.001). In order to confirm whether the P^®B diet also modified the cytokine expression levels in immune cells, primary cell cultures of splenocytes (Figure 3B–D) and peritoneal macrophages (Figure 3E,F) were carried out. In order to obtain detectable cytokine expression levels, cell cultures from CTR and LPS mice (LPS + SD and LPS + P^®B) were stimulated in vitro (splenocytes with Concanavalin A (ConA), and PEC-isolated macrophages with LPS).

As cytokine expression levels did not differ significantly between CTR + SD vs. CTR + P^®B groups, CTR represents the aggregate mean of all control mice, regardless of diet. In vitro stimulation of splenocytes obtained from LPS + SD mice induced a significant upregulation of the expression levels of the T helper 1 cytokines (i.e., Interferon (IFN)γ and IL-2; Figure 3B,C) as well as of the anti-inflammatory cytokine IL-10 (Figure 3D) (CTR vs. LPS + SD, p < 0.001). In macrophage cultures from LPS + SD mice, we observed a significant increase in IL-1β and TNFα expression (Figure 3E,F; CTR vs. LPS + SD, p < 0.001). The overexpression of cytokines both in splenocyte and macrophage cultures from LPS + P^®B mice was significantly reduced. Indeed, IL-2 and IL-10 expression levels (in splenocyte-derived culture, Figure 3C,D) of LPS + P^®B mice were not statistically different from CTR mouse levels (CTR vs. LPS+ P^®B, p > 0.05). Furthermore, in LPS + P^®B mice, even when in vitro stimulation led to an increase in the expression levels of IFNy (in splenocyte-derived culture, Figure 3B) and IL-1β and TNFα (in macrophage-derived culture, Figure 3E,F) compared to CTR mice (p < 0.001), such levels remained significantly lower compared to those of LPS + SD mice (LPS + SD vs. LPS + P^®B, p < 0.01).

2.4. Pain-Like Behaviour and Pro-inflammatory Cytokine mRNA Levels in Nervous Tissues

After the third and final LPS injection, pain-like behaviour was assessed. Both LPS + SD and LPS + P^®B mice showed a decrease in response thresholds to stimuli, indicating the presence of mechanical allodynia, assessed by the von Frey test (Figure 4A); thermal hyperalgesia, assessed with the plantar test (Figure 4B); and reduction in nociceptive thresholds (tail flick test, Figure 4C) (p < 0.001). Consistently, in all behavioural tests, LPS + P^®B mice exhibited significantly less allodynia, hyperalgesia and altered nociception compared to LPS + SD mice (p < 0.01).

To evaluate whether chronic feeding with P^®B was able to reduce not only sensory hypersensitivity but also LPS-induced alteration in the peripheral and central nervous system, pro-inflammatory cytokines particularly involved in pain were evaluated in key stations for pain transmission and modulation. Figure 5 reports the expression levels of IL-1β, IL-6 and TNFα evaluated in the sciatic nerves (Figure 5A–C), DRG (Figure 5D–F) and spinal cord (Figure 5G–I). In LPS + SD mice, a significant pro-inflammatory cytokine upregulation in both the peripheral (sciatic nerves and DRG) and central nervous system (spinal cord) was present (CTR vs. LPS + SD, p < 0.001). On the contrary, in LPS + P^®B mice, we did not observe a significant increase in pro-inflammatory cytokines in the sciatic nerves (Figure 5A–C; CTR vs. LPS + P^®B, p > 0.05), while in DRG, significant overexpression was present only for IL -1β (Figure 5D; CTR vs. LPS + P^®B, p < 0.001) and TNFα (Figure 5F; CTR vs. LPS + P^®B, p < 0.01). In the spinal cord (Figure 5G–I), all pro-inflammatory cytokines were upregulated compared to CTRs (CTR vs. LPS + P^®B, p < 0.001). It is noteworthy that in LPS + P^®B mice, even when an upregulation of pro-inflammatory cytokines was present, their levels remained significantly lower than those in LPS + SD mice, both in DRG (p < 0.05) and the spinal cord (p < 0.001).

3. Discussion

Our work demonstrates how inflammation and its behavioural comorbidities can be significantly ameliorated by a high-fibre legume sourdough bread diet in a murine LPS model.

The effects of P^®B may be explained, at least in part, by mechanisms involving the gut–brain axis, which plays a key role in the modulation of immune and neuroinflammatory responses. Indeed, the bidirectional links between the gut/microbiome system and the brain systems via the vagus nerve and various endocrine signals are well documented [16,43,75,76,77,78,79,80]. In this context, it is important to underline the iterative process prompted by customer feedback used for the development of P^®B. As noted above, the manufacturer conducted an in-depth inquiry in which customers who had spontaneously reported beneficial effects were asked to elaborate on their “holistic experience” of consuming P^®B, with emphasis on post-ingestive feelings of reward and associated internal sensations [44], which are linked to the unconscious affective responses to food [78,81]. The additional feedback obtained helped inform decisions regarding ingredient selection and processing methods, with the aim of optimizing not only texture and taste but also the overall consumer experience. This approach, grounded in fundamental aspects of the response of humans to food, may have contributed to selecting and improving the functional properties of P^®B, and although empirical and heuristic in nature, and therefore not rigorously scientific, it nonetheless applied some elements of product co-creation [82] and standardized affective food testing [83], which are employed in contemporary research on food preferences, hedonic evaluation and sensory experience [83,84,85].

The sourdough starter was generated de novo by the P^®B manufacturer with the traditional multistep procedure, promoting the growth of naturally occurring lactic acid bacteria (LAB) and yeasts [45,46,47,86,87,88]. The starter was propagated for several years with wheat flour using the traditional backslopping method [45,46,62,88], providing stable and reproducible fermentation performance in terms of leavening, rheological features and palatability enhancement [46,89]. The properties of breads obtained with this starter in terms of scents, flavours, acidity, shelf life, etc. are consistent with the expected presence of LAB and yeasts. However, the precise microbial composition of different sourdough starters can differ significantly, depending on a large number of factors that impact its ecology [47,87]. These include technological parameters (flour/water and starter/dough ratio, NaCl dosage, temperature, amount of oxygen incorporated during kneading, and properties of the tap water) and parameters that are impossible or difficult to control (microorganisms and properties in the ingredients, environment, etc.) [34,46,62,86,88]. A final and important step of selection of the sourdough microbiota occurs when the starter is presented with the other components of the final bread dough [87]. The ingredients of P^®B are numerous, some of them unusual, and each may contribute different functional properties; therefore, the overall recipe is different compared to other sourdough breads, and the different nutritional/ecological environment of the fermenting dough may enable the selection of an unusual or even unique microbiota [47]. Given the broad complexity of this important topic, the microbial composition of P^®B dough appeared outside the scope of the current investigation, and it has been reserved for future studies.

Table 1. Nutritional values of Primus^® bread compared to the standard diet. Percentages are normalized to the dry matter content of both products.

Nutritional Values	Primus® (100 g)	Primus®	Standard Diet V1534
Energy	876 KJ /207 kcal	387 kcal	349 kcal
Fats (g)	1.6	2.9%	3.8%
of which saturated (g)	0.4
Carbohydrates (g)	30	55.1%	47.0%
of which sugars (g)	0.9	1.7%	5.4%
Fibre (g)	4.4	8.1%	5.6%
Protein (g)	16	29.4%	21.7%
Sodium (g)	0.44	0.8%	0.3%

presents the nutritional composition of the two food treatments compared in this study, expressed in percentage of dry matter. Fat and sugars are 31% and 317% higher in the SD compared to P^®B, while carbohydrates and proteins are, respectively, 17% and 35% higher in P^®B vs. the SD; energy is 11% higher in P^®B. The differences in main nutritional components are within the range considered adequate for the C57BL/6J mice strain used in our study [90], and we believe that the component differences per se are unlikely to explain the differential effects described in this study. The SD and P^®B are also broadly comparable in terms of the categories of their raw ingredients: both consist of a similar combination of grains and legumes/legume-derived products. As is well known, this combination improves the nutrient value in terms of essential amino acid composition [5,28]. Consistent with the good nutritional profile of both diets, during the experimental protocol, the switch to the P^®B diet did not interfere with the normal growth curve, which was identical in both groups up until the LPS treatment. Animals fed P^®B had a very positive hedonic response to the new food, suggested by the significant increase in food intake on the first day after the change in diet. The increased intake was not caused by energetic demand, as the energy content of P^®B is higher than that of the SD. This novelty effect was the opposite of what is normally observed in rats and mice (including in the C57BL/6J strain) and in other mammals: a significant change in dietary offer normally causes taste neophobia and decreased intake when a new food is presented, even a highly palatable one with a different taste and ingredients. This is a reproducible phenomenon known as hyponeophagia, which is widely utilized in behavioural assays for anxiety and depression [91,92]. The feeding behaviour observed in response to P^®B was contrary to this typical hyponeophagic response. The reversal may be attributed to a combined effect of stimulus familiarity (as the ingredients share the same basic taste category as the SD) [93] and the apparent greater palatability of P^®B relative to the SD.

LPS administration induced a rapid and significant reduction in body weight in both groups; however, this effect was significantly attenuated in LPS-P^®B mice, consistent with a comparatively attenuated pathological state, in comparison with LPS-SD mice [94]. Likewise, differences in nociceptive behavioural responses between P^®B and the SD developed only after LPS injection.

The P^®B diet largely attenuated the systemic inflammatory response and the increase in pro-inflammatory cytokine mRNA in nervous tissues elicited by repeated LPS administration, reduced splenomegaly and prevented the expected peripheral cytokine surge, with a coherent and substantial effect on all parameters tested. Only modest increases in IL-1β, IL-12p40 and TNFα were detected, rather than the broad pro-inflammatory profile observed in LPS-SD mice. Correspondingly, the anti-inflammatory cytokines were higher in P^®B-LPS animals compared to the respective controls. In the spinal cord, the protective effect was only partial, with cytokine expression still elevated but at significantly lower levels compared to the respective controls. P^®B supplementation alone did not alter any basal inflammatory parameters relative to SD-fed mice. This observation suggests that P^®B does not show any immunosuppressant properties in healthy mice.

Ex vivo stimulation of splenocytes and peritoneal macrophages confirmed the protective profile, as P^®B-fed LPS animals displayed significantly lower expression of pro-inflammatory cytokines than SD-LPS mice. The attenuation of splenomegaly and the reduced responsiveness of splenocytes and macrophages confirm a lower level of systemic immune activation, which seems to be associated with the reduced weight loss as well as to the attenuated nociceptive sensitization. In SD-fed animals instead, all parameters were consistent with the well-established cascade in which systemic LPS induces cytokine release, activates peripheral immune organs, disrupts the blood–brain barrier and ultimately triggers central alteration of pro-inflammatory cytokine expression and pain hypersensitivity [73,95].

The protection of behavioural parameters observed in P^®B-fed mice is particularly remarkable. LPS-induced hyperalgesia is mediated by both direct actions of pro-inflammatory cytokines on nociceptors and by central sensitization processes involving glial activation [70]. The blunted cytokine responses in the sciatic nerves and DRG provide a plausible mechanistic link to the reduced hypersensitivity in LPS-P^®B mice, suggesting that diet can modulate the interface between immune activation and nociceptor sensitization [12]. The incomplete protection in the spinal cord may reflect the potency of repeated doses of LPS in directly activating central pathways, including microglia, via Toll like Receptor-4 and related mechanisms, as previously shown [73].

In nervous tissues, we measured the expression of genes of pro-inflammatory cytokines but not protein expression nor the expression of cellular markers of glial activation; therefore, these readouts may not be enough to fully characterize the presence of neuroinflammation. However, in other murine pain models in which neuroinflammatory markers were assessed both as mRNA and protein, we consistently observed a good correspondence between transcriptional and protein changes (e.g., [96,97]).

It is not possible to unequivocally identify the specific component(s) of the P^®B diet or the mechanisms distinguishing it from the SD that are responsible for the observed effects, because the compositional information for both products is incomplete: exact ingredient lists (SD) and their exact relative proportions (P^®B) are proprietary to the manufacturers. This represents an inherent constraint for our study in isolating the specific drivers of the observed effects. Nevertheless, the available data allow several potentially relevant differences to be highlighted. First, the main ingredient of P^®B is spelt, a grain that is not included in the SD and which manifests enhanced functional properties upon fermentation (discussed below). Barley, rye, and oats are present in both diets. Previous studies have reported that barley, rye, and oats exert moderate anti-inflammatory effects in human subjects with pre-existing health complications, likely mediated by β-glucans and favourable modulation of gut microbiota composition, but show little or no effect in healthy individuals [98,99,100]. Such properties would contribute partially to the differential actions of P^®B only if these grains were present at significantly higher levels in P^®B than in the SD, which is unlikely. Second, P^®B contains substantially more fibre (+45%), derived from several sources, compared with the SD, which act as prebiotics in the gut (see below). Third, P^®B ingredients derived from other ancient grains (Oroset and Virgo soft wheats and Senatore Cappelli durum wheat—see Methods) are unlikely to be present in the SD, as they are niche products not typically used in industrial animal feeds. The “Senatore Cappelli” cultivar of durum wheat, an ancient grain variety derived from a pure line selection of a Tunisian Jeann Rhetifah ecotype, contributes polyphenols and flavonoids of greater diversity compared to wheat, antioxidant activity and other beneficial effects [101,102,103,104], including the improvement of inflammation-related symptoms in patients with non-celiac gluten sensitivity [104,105]. Fourth, there are some ingredients of P^®B (unconventional in Italian sourdough breads) whose presence is unknown in the SD that may be relevant for P^®B effects. Psyllium husk and malt are typical bakery ingredients unlikely to be present in the SD. Malt can significantly modulate the sourdough ecology and metabolisms of the sourdough microbiota [106]; millet contains a range of bioactive compounds [107], including polyphenols, flavonoids and several others [108], some of which are associated with anti-inflammatory and antioxidant activities in both in vitro and dietary contexts [109,110]; linseed (flaxseed) is one of the richest dietary sources of lignans [111], fibres and other metabolites reviewed for their anti-inflammatory and antioxidant effects, in particular the ω-3 fatty acid α-linolenic acid (ALA) [112,113]; wheat germ contains phenolic acids, tocopherols, lignans, lipids and vitamins, all of which are able to contribute anti-inflammatory effects [114,115,116,117,118,119], as well as gut flora-regulating peptides released during fermentation [117]. Fifth, many ingredients of P^®B are certified organic, while the SD ingredients are not. Small to moderate differences have been shown between organic and conventional crops in antioxidant activity, increased mineral availability, phenolic content and other beneficial components in spelt and other cereals [120], but, in particular, sourdough starters fed with organic cereals showed significantly increased complexity in the microbiota [121]. Therefore, the organic origin of the main P^®B ingredients should be considered as one of the relevant differences between P^®B and SD. Sixth, P^®B has a higher protein content than the SD, because purified proteins from several different pulses and cereals were added; the use of protein isolates instead of whole legume flours has the advantage of avoiding the various antinutritional factors present in whole pulses while providing different proteins with specific properties [6,25,29,31,32,37,38,40,46,47,50,53,64,117,122,123].

Finally, a fundamental difference lies in the processing of P^®B ingredients, which undergo fermentation and baking, relative to the untreated ingredients in the SD. This processing disparity is likely the primary driver of the substantial changes of the P^®B diet vs. SD observed in the LPS mouse model. Sourdough-fermented spelt (SFS) as P^®B’s main ingredient can contribute significant anti-inflammatory properties, thanks to the high flavonoid and phenolic content, high antioxidant activity, angiotensin-converting enzyme inhibitory effects, and other relevant anti-inflammatory effects detected in SFS in in vitro essays [25,26,124]. We acknowledge that the cumulative impact of multiple nutritional and processing variables represents a potential confounding factor. However, it is also possible that this complexity is instrumental, or even requisite, for the observed differential effects between the two diets. This holistic impact warrants further investigation to tease out specific versus synergistic mechanisms.

Sourdough breads are considered a valuable source of several prebiotic components (cellulose, resistant starch, arabinoxylans, and β-glucans), resulting from the fermentation of dietary fibres and proteins [6,31,32,38,47,50,87,123], some of which are attracting interest for their therapeutic potential [3,6,7,31,32,38,46,122]. Additional benefits of fermentation include the reduction in antinutrients; the enhanced bioavailability of minerals, vitamins, and free amino acids [10,54,125,126]; and the production of a diverse array of bioactive peptides with interesting properties (reviewed in [10]). FODMAPs (Fermentable Oligo-, Di-, and Mono-saccharides and Polyols) are the non-digestible, rapidly fermentable short-chain carbohydrates present in cereals, pulses and other vegetable foods [127]. High concentrations of these constituents can cause digestive distress, particularly in patients with irritable bowel disease and wheat sensitivity [4,7]. Sourdough fermentation decreases the levels of FODMAPs by 50–80% without affecting the beneficial, slowly fermenting dietary fibre, making fermented breads more tolerable for such patients [4,7,128]. A low-FODMAP environment in the gut promotes the growth of specific families of beneficial lactic bacteria (Bifidobacterium and many others), resulting in a significant modulation of the microbiota [129,130,131,132]. Presumably, although not measured, the level of FODMAPs in P^®B are low, thanks to the long fermentation procedure.

The prebiotic fibres (higher in P^®B than SD, as said) feed the activity of the gut microbiota, leading to the synthesis of short-chain fatty acids (SCFAs). SCFAs synthesis also occurs during dough fermentation; as a result, sourdough breads, particularly the ones (like P^®B) prepared with an extended fermentation procedure, already contain significant amounts of SCFAs, which therefore act as postbiotics after ingestion [34,131,133]. As well as exerting several systemic effects on cognitive functions and behaviour, SCFAs modulate local and systemic inflammation in multiple highly relevant ways for the present study [133,134,135]. Butyrate in particular is able to increase expression of IL-10 and IL-23 and decrease IL-12 and IFNγ expression in dendritic cells, downregulate key components of the antigen presentation apparatus and major histocompatibility complex class II [136,137], regulate autophagy, inhibit histone acetylase, and act as ligands of free fatty acid receptor 2 (FFAR2) and 3 (GPR43, GPR41) in macrophages, neutrophils and monocytes [133]. In peripheral blood mononuclear cells, SCFAs downregulate monocyte chemotactic protein-1 and the synthesis of TNF, IFN-γ and IL-10 in response to treatment with LPS [138]. These and several other functions of SCFAs may contribute significantly to the immunomodulatory effectiveness of P^®B vs. the SD in LPS-inflamed mice [9,15,138].

In conclusion, our results demonstrate that, in our experimental model, supplementation with the functional food P^®B appears to substantially mitigate the LPS-induced inflammatory cascade, particularly in its peripheral components. Our data, although primarily associative, agree with the anecdotical observation of a correlation between the beneficial effects reported by human subjects with inflammatory conditions and the incorporation of P^®B in their daily dietary habits. The notable efficacy observed in an in vivo model of systemic and central inflammation is likely the result of the combined multifactorial contributions, possibly synergistic, of several elements (prebiotics, postbiotics, bioactive compounds, etc.) and on their interaction with the gut microbiota. It remains to be determined if other sourdough breads with different ingredients and different fermentation protocols will be able to reproduce similar anti-inflammatory effects in the same model. We acknowledge that further studies are needed in order to translate the results of our study into the insertion of this functional food in the human diet for clinical purposes.

4. Materials and Methods

4.1. Animals

A total of 28 male 9-week-old C57BL/6J mice (Charles River Laboratories, Calco, Italy) were used in this study. Multiple animals were housed per cage (Macrolon type II cages equipped with sawdust bedding, nesting and environmental enrichment) with a 12 h light/dark cycle, temperature of 22 ± 2 °C and humidity of 55 ± 10%. The animals were subjected to a different diet depending on the experimental groups (see Section 4.3, Experimental Protocols) and with access to tap water ad libitum. Before the beginning of the experimental procedures, all mice were acclimated and habituated to handling for 1 week. All procedures were conducted in accordance with European Directive 2010/63/EU and approved by the Animal Care and Use Committee of the Italian Ministry of Health (authorization number 470/2016-PR given to PS on 16 May 2016).

4.2. Bread

Primus^® bread (P^®B) is an artisanal product, distributed nationally in Italy by the small family-run company Antico Forno by Santolin Sandro srl, Carpi, Modena, Italy. The natural yeast (sourdough starter) used for ingredient fermentation/leavening is a strain of wild origin, and although it is proprietary, the company agrees to make it freely available upon request. The fermentation is an extended (24 h) multistep process controlled exclusively through the optimization of thermal parameters and the duration and intensity of mechanical treatments. The current formulation of P^®B was refined through multiple cycles involving extensive testing of diverse combinations of cereals and proteins from grains and legumes, with emphasis on the diversification of raw materials selected for local availability, for their nutritional and functional properties, for customer satisfaction (see the Introduction) and for sustainability. The iterative design process focused on achieving optimal palatability and maximizing customer satisfaction (further details in the Discussion). The final ingredients chosen were wholemeal flours of spelt *, barley *, rye *, oats *, and soft wheats * (“Oroset” [139] and “Virgo” [11]); sourdough starter (wheat); isolated proteins from mixed legumes and cereals (soya, soft wheat, peas, lupin beans, and oats); durum wheat bran * (“Senatore Cappelli”); spelt bran*; durum wheat semolina * (“Senatore Cappelli”); flours of millet *; linseeds *; toasted soy *; soft wheat germ *; whole sea salt (“Cervia”); malt flour; and psyllium husk powder (* certified organic).

The ingredients of the standard diet declared on the ssniff^® (Ssniff Spezialdiäten GmbH, Soest, Germany) datasheet are “grain and grain by-products, oil seed products, minerals, vegetable oils, vitamins, trace elements”. Although the exact composition is proprietary, after enquiry we obtained that the ingredients “…are mainly cereal grains, i.e., wheat and wheat midds, barley and oat bran, soybean meal (solvent extracted), full-fat soybeans (heat treated/toasted)”. The nutritional values of P^®B and of the standard diet are shown in Table 1.

4.3. Experimental Protocols

In Figure 6 the schematic experimental schedule is reported. After 1 week of housing with the standard pellet diet (SD), on day 0, mice (n = 28) were randomized using the GraphPad Prism Quickcalcs free online tool [140] in two experimental groups: animals fed with standard dry pellets (SD group) (n = 14, diet V1534—Ssniff Spezialdiäten GmbH, Soest, Germany) and animals fed with Primus^® bread dried at low temperature (P^®B group) (n = 14, for ingredients see Section 4.2 and Table 1). During the experimental protocol, a new batch of P^®B was supplied fresh every day and the old one discarded after weighting. On day 14, both experimental groups were randomized (by GraphPad online source) into two further groups: animals treated with 3 mg/kg of lipopolysaccharide (LPS, Merck, Darmstadt, Germany) or vehicle (10 µL/g) intraperitoneally (i.p.) for 3 consecutive days (n = 7/group) [95,141,142]. LPS was freshly prepared by dissolving it in saline solution. Throughout the experimental schedule, mouse body weight and food consumption were monitored daily. On day 16, 2 h after the third and last LPS or vehicle administration, pain-like behaviours were tested, followed by animal euthanasia by decapitation. After sacrifice, blood, peritoneal macrophages, spleen (for splenocyte isolation), sciatic nerves, dorsal root ganglia (DRG, L4–L6 portion) and spinal cord (L4–L6 portion) were separately collected for subsequent biochemical evaluations.

4.4. Plasma Collection and Analysis

The whole blood from the carotid artery was collected in tubes containing the anticoagulant agent heparin (5000 I.U./mL, Pfizer, Italy). The tubes were kept at 4 °C until centrifugation (15 min, 1500 g at 4 °C). Once the separation of the whole blood was achieved, plasma was collected and stored at −20 °C, Plasma was used for evaluation of a 32-analyte panel (i.e., Eotaxin, G-CSF, GM-CSF, IFNγ, IL-1α, IL-1β, IL-2, IL-3, IL-4, IL-5, IL-6, IL-7, IL-9, IL-10, IL-12p40, IL-12p70, IL-13, IL-15, IL-17A, IP-10, KC, LIF, LIX, MCP-1, M-CSF, MIG, MIP-1α, MIP-1β, MIP-2, RANTES, TNFα and VEGF-A) via an external ELISA analysis service (Mouse Cytokine/Chemokine Array Discovery Assay^® 32-Plex (MD32)—Eve Technologies Corporation, Calgary, Canada).

4.5. Cell Sampling and Stimulation for Cytokine mRNA Level Evaluations

Macrophages: Macrophages were isolated from peritoneal cells (PECs) as previously described [23,24]. Briefly, after the sacrifice, the mouse peritoneum was washed with 10 mL of RPMI 1640 medium (Sigma-Aldrich, Milan, Italy) supplemented with 10% FBS (Gibco, Thermofisher Scientific, Waltham, USA). An aliquot of the gathered sample was used to check cell viability (trypan blue (Sigma-Aldrich, Milan, Italy)) and nuclei morphology (Turk’s solution (Sigma-Aldrich, Milan, Italy)) as well as for cell counting. PECs were dispensed into 24-well culture plates at a final concentration of 10⁶/mL per well. Macrophage purification and isolation were performed by 2 h plate adhesion (macrophage purity around 90%). Subsequently, to stimulate cytokine production, lipopolysaccharide (LPS, Sigma-Aldrich, Milan, Italy) was added to the adhered macrophages at a concentration of 1 µg/well in a final volume of 1 mL/well of complete culture media (RPMI 1640 supplemented with 10% FBS, 2% glutamine, 1% antibiotics and 0.1% 2-mercaptoethanol). The stimulus was not added to control wells. After 24 h of incubation (37 °C and 5% CO₂), adhered cells were detached from the plate by adding 500 µL/well of TRIzol™ Reagent (Invitrogen, Carlsbad, USA) and stored at −80 °C for biochemical evaluations (RT-qPCR).

Splenocytes: After sacrifice, the spleen of each mouse was rapidly removed and weighed under aseptic conditions. Splenocytes were collected through an incision on the spleen cuticle, as previously described [143]. In brief, isolated cells were placed in 24-well plates at the final concentration of 4 × 10⁶ cells/mL in complete culture media and incubated at 37 °C in 5% CO₂ with or without 10 μg/mL Concanavalin A (ConA), for stimulation of Thelper1 and Thelper2 cytokines. The stimulus was added to the cell cultures in a final volume of 1 mL/well. Cell pellets were collected after 24 (for IFNy and IL-2 mRNA level assessments) and 48 h (for IL-10 mRNA level assessments) of culture and stored at −80 °C for biochemical analysis (qRT-PCR).

4.6. Collection of Nervous Tissues

Sciatic nerves: A small vertical incision was made along each thigh, and the underlying musculature was separated to expose the sciatic nerve. Subsequently, this nerve was lifted and removed by making a cut at the proximal and distal ends. For each mouse, the right and left sciatic nerves (approximately 1 cm each) were collected.

Spinal cord: The entire spinal column (from the sacrum to the atlanto-occipital joint) was dissected. Following from the caudal spinal canal and proceeding carefully along the sagittal plane, one vertebra at a time (keeping to the vertebrae centre) was cut in the cranial direction until the spinal column was semi-open. The spinal cord was thus exposed and L4–L6 was removed (by counting the vertebrae in the caudal direction, from thoracic 1 (T1).

Dorsal root ganglia: The vertebral column, deprived of the spinal cord, was cut in half along the sagittal plane and the spinal canal was kept facing upwards. The DRG L4–L6 were dissected from the intervertebral foramina by drilling the meninges under a stereomicroscope (Olympus, Milan, Italy). The procedure was also repeated on the other half of the vertebral column. For each mouse, 6 DRGs were collected.

All nervous tissues were immediately frozen in liquid nitrogen and stored at −80 °C for subsequent biochemical analyses (RT-qPCR).

4.7. RT-qPCR

The RNA was isolated from macrophages, splenocytes, sciatic nerves, DRG and spinal cord using TRIzol™ (Invitrogen, USA) reagent and reverse transcribed to cDNA with the LunaScript^® RT SuperMix Kit (BioLabs, Altrincham, UK). Marker expression levels were analyzed by Real-Time PCR (QuantStudio 5™—Thermofisher Scientific, Waltham, USA) using Luna^® Universal qPCR Master Mix (BioLabs, Altrincham, UK) and TaqMan probes (Thermofisher Scientific, Waltham, USA) specific to the genes under study (IL-1β, Mm00434228_m1; TNF-α, Mm00443258_m1; IFNγ, Mm01168134_m1; IL-2, Mm00434256_m1; and IL-10, Mm01288386_m1) according to the manufacturers’ instructions. Each sample was run in triplicate alongside non-template controls (NTCs). The mRNA levels of the interest genes were normalized to the endogenous gene (GAPDH—Mm99999915_g1) and expressed as 2−^ΔΔCt relative to the control group (control group included all CTR animals, regardless of diet).

4.8. Pain-Like Behaviour

On day 16, approximately 2 h after the last LPS injection (around 10.00 am), the presence of pain-like responses was assessed by trained researchers blinded to the treatment/diet. Before the assessments, mice were familiarized with the new environment (behavioural room) for approximately 30 min.

Mechanical allodynia: It was evaluated using a Dynamic Plantar Aesthesiometer (von Frey test, model: 37550—Ugo Basile, Gemonio, Italy). Briefly, the animals were placed in a modular plexiglass enclosure (1 mouse per module) positioned over a wire mesh. A blunt probe (von Frey filament; 0.5 cm diameter) was used to apply a mechanical stimulus to the mid-plantar surface of the animal’s hind paws. The force was steadily increased (up to 10 g over 10 s) until the animal withdrew its paw in response to the painful stimulation. The instrument automatically measured and displayed the corresponding force in grams, recorded as the paw withdrawal threshold (PWT) [144].

Thermal Hyperalgesia: It was evaluated using the plantar test (model: 37570—Ugo Basile, Gemonio, Italy). Briefly, mice were placed in a modular plexiglass enclosure (1 mouse per module) over an opaque glass plate. A constant thermal stimulus, generated by a focused light source (beam 0.5 cm and intensity 20 I.R.) was applied through a glass panel to the plantar surface of the mice’s hind paws. The time taken for the mouse to withdraw the stimulated paw (the response to the stimulus) was recorded in seconds and defined as the paw withdrawal latency (PWL). A 22 s cut-off was used to prevent tissue damage [145].

Nociceptive thresholds: The nociceptive threshold to heat stimulus on mouse tails was assessed using the tail flick test (instrument model: DS20 Socrel—Ugo Basile, Gemonio, Italy). Briefly, a small section of the mouse’s tail (i.e., approximately 2 cm above the tail tip) was placed under a radiant heat source. Pain sensitivity was measured by tail flick latency (in seconds), defined as the time from the onset of radiant heat to tail withdrawal (TWL, tail withdrawal latency); cut-off was 10 s [146]. For paw withdrawal thresholds and latency (PWT and PWL), three measurements were taken for each hind paw of every mouse. The values obtained from both hind paws were averaged and the values obtained were used for statistical analysis.

4.9. Statistical Analysis

The statistical analysis was performed using GraphPad Prism 9 (San Diego, CA, USA). Normality and equal variances were checked before choosing statistical tests. Data of body weight and food intake were analyzed using Two-way ANOVA with Bonferroni’s post hoc test, whereas data of pain-like behaviour and biochemical parameters were analyzed using One-way ANOVA followed by Bonferroni’s test. Data indicates the mean ± SEM of 14 or 7 mice/group (for detail, see figure legends). For all analyses, differences were considered significant at p ≤ 0.05.