4.2. Possibility of Phage Therapy in the Treatment of CNS-Related Bacterial Infections
4.2.噬菌体疗法治疗中枢神经系统相关细菌感染的可能性
With dramatically increasing antibiotic resistance, alternative therapeutic strategies are critically needed. Phage therapy (PT) has been considered a vital part of the solution [152,153]. Bacteriophages offer several advantages, including (i) high specificity, allowing selective targeting of pathogenic bacterial strains with no negative impact on the natural microbiota; (ii) self-replication at the side of infection; (iii) the ability to kill antibiotic-resistant bacterial strains; (iv) a lower risk of spreading resistance mechanisms; (v) coevolution with the bacterial host; (vi) safety even for immunocompromised patients; and (vii) little to no side effects [73]. Moreover, the majority of the antibiotics approved over the past decades represent already known classes of antimicrobials, with only a fraction having novel chemical structures [154].
随着抗生素耐药性的急剧增加，迫切需要替代治疗策略。噬菌体疗法（PT）被认为是解决方案的重要组成部分[ 152, 153]。噬菌体具有多种优势，包括：(i) 特异性强，可选择性地靶向病原菌菌株，对自然微生物群没有负面影响；(ii) 在感染侧可自我复制；(iii) 能够杀死耐抗生素的细菌菌株；(iv) 耐药性机制扩散的风险较低；(v) 与细菌宿主共同进化；(vi) 即使对免疫力低下的患者也很安全；(vii) 几乎没有副作用[73]。此外，过去几十年中批准的大多数抗生素都是已知的抗菌素类别，只有一小部分具有新的化学结构[154]。
The very first report on using bacteriophages in the treatment of brain-related bacterial diseases dates back to 1943 when Dubos et al. administrated phages specific to Shigella dysenteriae intraperitoneally (ip) in a mouse model [155]. What is more, the applied phages were detected in the brains of the animals to which they were administered. After phage application increased, phage titer was observed in infected mice. Interestingly, out of eight mice treated with high bacteriophage titer (109 PFU/animal), six animals survived, whereas out of eight pup rats treated with lower phage titer (105 PFU/animal) only two survived. The next phage breakthrough came in 1982 when Smith and Huggins conducted an experiment using bacteriophages in the treatment of E. coli producing colicin V, which causes septicemia and meningitis [156]. They demonstrated that Phage R was highly active in vitro, which was reflected in vivo (when 3 × 104 phage particles given intramuscularly or 3 × 103 given intravenously were sufficient) to cure mice administered a potentially lethal intramuscular dose of E. coli. The authors also demonstrated that although phage R-resistant mutants are relatively common in laboratory cultures, they were only found in a few phage-treated mice, and even then they were confined to the inoculation site and outnumbered by phage-sensitive cells. The efficacy of PT vs. antibiotic treatment was compared and showed that one dose of phage R was at least equivalent to multiple doses of streptomycin and more effective than multiple doses of tetracycline, ampicillin, chloramphenicol, or trimethoprim plus sulphafurazole in treating mice infected intramuscularly or intracerebrally with E. coli. The therapeutic success of PT was due to high in vivo activity and the failure of phage-resistant mutants to proliferate during treatment [156]. In 1998, Barrow et al. investigated E. coli causing septicemia and meningitis-like infection and newly isolated phage R (procured from sewage water) selective for K1 antigen-possessing bacterial strains [157]. This antigen is a capsular polysaccharide, associated with a highly virulent nature, and is often found in pathogenic E. coli [158]. When chickens were intramuscularly infected with 106 CFU or intracranially inoculated with 103 CFU of E. coli and simultaneously intramuscularly treated with 106 or 108 PFU of phage R (in the case intramuscular or intracranial bacterial administration, respectively), no morbidity or mortality of the animals was observed [157]. The authors administered lower doses of bacteriophage preparation into the chickens and found that in the case of the intramuscular bacterial inoculation, the application of 102 PFU caused some protection against bacteria, but the effect was not statistically significant. However, with the intracranial bacterial administration, doses lower than 108 PFU were not successful. Interestingly, the authors also observed that newly hatched chickens could acquire some protection against pathogens with lower doses of bacteriophages (106 PFU) compared to older chickens (108 PFU). They also tested the effectiveness of PT on colostrum-deprived calves infected orally and demonstrated that in calves treated intramuscularly with phage R, bacterial infection was controlled, yet the infection could not be prevented altogether. Administration of PT delayed the onset of signs of E. coli bacteremia.
最早关于使用噬菌体治疗脑部相关细菌疾病的报道可追溯到 1943 年，当时 Dubos 等人在小鼠模型中腹腔注射了针对痢疾志贺氏菌的噬菌体[155]。更重要的是，施用噬菌体的动物大脑中检测到了噬菌体。噬菌体应用增加后，在受感染的小鼠体内观察到了噬菌体滴度。有趣的是，在使用高滴度噬菌体（10 9 PFU/只）治疗的八只小鼠中，有六只存活下来，而在使用低滴度噬菌体（10 5 PFU/只）治疗的八只幼鼠中，只有两只存活下来。噬菌体的下一个突破出现在 1982 年，当时史密斯和哈金斯利用噬菌体进行了一项治疗大肠杆菌的实验，这种大肠杆菌会产生可导致败血症和脑膜炎的大肠杆菌毒素 V[ 156]。他们证明，噬菌体 R 在体外具有很高的活性，这在体内也得到了反映（肌肉注射 3 × 10 4 噬菌体颗粒或静脉注射 3 × 10 3 就足以治愈肌肉注射可能致命的大肠杆菌剂量的小鼠）。作者还证明，虽然噬菌体 R 抗性突变体在实验室培养物中比较常见，但它们只出现在少数经过噬菌体处理的小鼠体内，即便如此，它们也仅限于接种部位，而且数量比噬菌体敏感细胞要多。在治疗肌肉或脑内感染大肠杆菌的小鼠时，比较了噬菌体治疗与抗生素治疗的疗效，结果表明，一剂噬菌体 R 至少相当于多剂链霉素，而且比多剂四环素、氨苄西林、氯霉素或三甲双胍加磺胺嘧啶更有效。PT 的治疗成功是由于其体内活性高，而且在治疗过程中抗噬菌体的突变体不会增殖[ 156]。1998 年，Barrow 等人研究了引起败血症和类似脑膜炎感染的大肠杆菌，以及新分离出的噬菌体 R（从污水中提取）对含有 K1 抗原的细菌菌株的选择性[157]。这种抗原是一种荚膜多糖，具有很强的毒性，通常存在于致病性大肠杆菌中[ 158]。当鸡肌肉注射 10 6 CFU 或颅内注射 10 6 CFU 时，大肠埃希氏菌的抗原就会出现。CFU 或颅内接种 10 3 大肠杆菌和 10 3 大肠杆菌的菌株。CFU 的大肠杆菌，同时肌肉注射 10 6 或 10 8 PFU的噬菌体R（分别用于肌肉注射或颅内细菌注射），未观察到动物发病或死亡[ 157]。 作者给鸡注射了较低剂量的噬菌体制剂，发现在肌肉细菌接种的情况下，注射 10 2 PFU 对细菌有一定的保护作用，但效果没有统计学意义。PFU 对细菌有一定的保护作用，但效果在统计学上并不显著。然而，在颅内注射细菌制剂时，低于 10 {{9} PFU 的剂量并不能成功杀灭细菌。｝PFU 的效果不佳。有趣的是，作者还观察到，与年龄较大的鸡（10 8 PFU）相比，新孵化的鸡在使用较低剂量的噬菌体（10 6 PFU）时也能获得一定的抗病原体保护。他们还测试了噬菌体对缺失初乳的小牛口服感染的有效性，结果表明，对小牛肌肉注射噬菌体 R 可控制细菌感染，但不能完全防止感染。给小牛注射噬菌体可延缓大肠杆菌菌血症症状的出现。
Poulliot et al. demonstrated that isolated lytic bacteriophages can neutralize E. coli strain S242, which causes sepsis and one of the most severe E. coli infections, fatal neonatal meningitis, with a mortality rate of up to 25% [30]. The mortality rate may increase as the infection itself may worsen if the pathogen acquires multi-drug resistance (MDR), and E. coli strains expressing acquired extended-spectrum beta-lactamases (ESBL), such as CTX-M-type enzymes [159], turned out to be especially lethal. The major one is CTX-M1, which is an enzyme with hydrolytic activity against cefotaxime. The EC200PP bacteriophage (isolated from environmental sewage samples in France) was characterized in vivo and ex vivo with regard to stability and pharmacokinetic properties in pup rats. The EC200PP was stable for at least 24 h in adult and pup rats’ serum, while in human serum phage titer decreased by 2 to 3 logs in the first 2 h of incubation and remained stable after that. As stated in the article, serum possesses phage-inhibitory activity, and this activity is not specific to EC200PP, but rather to all phages. When phage preparation was administrated, most of the phage virions were preferably located in the spleen and kidneys, while low titers were observed in urine and the CNS. Because E. coli S242 strain may induce sepsis and meningitis, both models were investigated. For the sepsis model, pup rats were injected with 104 CFU of S242 bacteria ip and then administered 108 PFU/mL of EC200PP subcutaneously 7 h or 24 h after the induction of infection. Interestingly, a 7 h post-infection intervention resulted in 100% recovery, while phage administration 24 h after infection resulted in only 50%. For meningitis models, two scenarios were investigated: with a lower dose of the S242 bacteria administered intrathecally, and with a higher inoculum of bacteria, corresponding to that normally encountered in human meningitis. The first case was based on administering 200 CFU per rat to cistera magna, which resulted in a titer of 106 CFU/mL in the CSF after 24 h. Pup rats were then treated with 108 PFU of EC200PP per pup rat administered ip 1h post-infection, which contributed to a 100% survival rate up to day 5. This also contributed to nondetectable bacterial levels in the CSF, and EC200PP titer of 4.5 +/− 0.2 log PFU/mL 24 h post-infection. Based on the highly successful initial results, a second scenario was then investigated, with a higher concentration of S242 bacteria (2 × 106 CFU) administered intrathecally, relevant to that observed in human meningitis, and then 108 PFU of EC200PP 1h post-infection. As a result, all rats survived to day 5 of the experiment. After 24 h, the CSF samples were positive for bacteria, but no signs of bacteremia were observed, and on day 5, in three out of five rats, the CSF was sterile with no signs of infection. However, when phage treatment was delayed to 2 or 3 h post-infection, the survival rate decreased dramatically [30].
Poulliot 等人证实，分离出的溶解性噬菌体可以中和大肠杆菌 S242 菌株，该菌株可导致败血症和最严重的大肠杆菌感染之一–致命性新生儿脑膜炎，死亡率高达 25%[ 30]。如果病原体获得多重耐药性（MDR），死亡率可能会随着感染本身的恶化而增加，而表达获得性广谱β-内酰胺酶（ESBL）（如 CTX-M 型酶）的大肠杆菌菌株[159]尤其致命。其中最主要的是 CTX-M1，它是一种对头孢他啶具有水解活性的酶。EC200 PP 噬菌体（从法国的环境污水样本中分离出来）在幼鼠体内和体外的稳定性和药代动力学特性得到了表征。EC200 PP 在成年大鼠和幼鼠血清中至少 24 小时内保持稳定，而在人血清中，噬菌体滴度在培养的头 2 小时内下降 2 到 3 个对数，之后保持稳定。正如文章所述，血清具有抑制噬菌体的活性，而且这种活性不是针对 EC200 PP 的，而是针对所有噬菌体的。，而是对所有噬菌体都有抑制作用。在服用噬菌体制剂时，大多数噬菌体病毒最好位于脾脏和肾脏，而在尿液和中枢神经系统中则观察到较低的滴度。由于大肠杆菌 S242 菌株可能诱发败血症和脑膜炎，因此对这两种模型进行了研究。在败血症模型中，给幼鼠静脉注射 10 4 CFU 的 S242 细菌。CFU的S242菌株，然后给幼鼠注射10 {{4}在诱导感染后7小时或24小时，皮下注射10 {{4} PFU/毫升的EC200 PP 。有趣的是，感染后 7 小时进行干预可使患者 100% 康复，而感染后 24 小时注射噬菌体仅可使患者 50% 康复。对于脑膜炎模型，研究了两种情况：一种是经皮下注射较低剂量的 S242 细菌，另一种是接种较多的细菌（相当于人类脑膜炎中通常遇到的接种量）。第一种情况是在每只大鼠的水泡中注射 200 CFU，结果在脑脊液中的滴度为 10 6 CFU/mL然后用 10 {{7}PFU 的 EC200 {{7} 处理幼鼠。8}} 的 EC200 {{7} PFU 对幼鼠进行治疗，感染后 1 小时ip给药，这使得幼鼠在第 5 天前的存活率达到 100%。这也导致脑脊液中检测不到细菌水平，感染后24小时EC200 PP 滴度为4.5 +/- 0.2 log PFU/mL。在初步结果非常成功的基础上，研究人员又研究了第二种情况，即鞘内注射更高浓度的 S242 细菌（2 × 10 6 CFU），这与在人类脑膜炎中观察到的细菌浓度相似，然后再注射 10 8 PFU 的 EC200 {{10} }。PFU的EC200 PP 感染后 1 小时。结果，所有大鼠都活到了实验的第5天。 24 小时后，脑脊液样本中的细菌呈阳性，但没有观察到菌血症的迹象，第 5 天，5 只大鼠中有 3 只的脑脊液无菌，没有感染迹象。然而，当噬菌体治疗延迟到感染后 2 或 3 小时时，存活率急剧下降[ 30]。
Bacteriophage K1F specific to E. coli strain EV36 also expressing the K1 capsule antigen was used in an in vitro model of bacterial neonatal meningitis in human cerebral microvascular endothelial cells (hCMEC), which are a part of the BBB [160,161,162,163]. The studies revealed that the K1F phage was internalized by phagocytosis upon contact with hCMEC cells, and then degraded through a constitutive pathway—via lysosomal degradation along with PAMP-LC3-dependent phagocytosis, which suggests that phages are recognized by Toll-Like Receptor 3 (TLR3) on mammalian cells. Importantly, the K1F phage did not induce an inflammatory response. The authors also reported temporarily decreased barrier resistance of hCMEC cells, which could facilitate the transition of immune cells across the endothelial vessel in vitro. Importantly, K1F can infect the EV36 strain intracellularly within hCMEC cells without inducing inflammation [29].
噬菌体 K1F 特异于同样表达 K1 胶囊抗原的大肠杆菌 EV36 菌株，被用于人脑微血管内皮细胞（hCMEC）细菌性新生儿脑膜炎的体外模型，而人脑微血管内皮细胞是 BBB 的一部分[ 160, 161, 162, 163]。研究发现，K1F 噬菌体与 hCMEC 细胞接触后通过吞噬作用被内化，然后通过构成性途径–溶酶体降解和 PAMP-LC3 依赖性吞噬作用–降解，这表明噬菌体能被哺乳动物细胞上的 Toll-Like Receptor 3（TLR3）识别。重要的是，K1F 噬菌体不会诱发炎症反应。作者还报告说，hCMEC 细胞的屏障阻力暂时下降，这可能会促进免疫细胞在体外穿过内皮血管。重要的是，K1F 可以在 hCMEC 细胞内感染 EV36 菌株，而不会诱发炎症[ 29]。
Apart from whole phage virions, bacteriophage endolysins were also recently documented to demonstrate great potential for treating severe cases of infections with gram-negative and gram-positive bacteria, such as E. coli, Salmonella spp., Pseudomonas aeruginosa, Pseudomonas putida, Shigella boydii, Shigella flexneri, Vibrio fischeri, Vibrio vulnificus, and many more [164,165,166,167,168]. Phage endolysins are subdivided into five families depending on the peptidoglycan disruption mechanism: glucosaminidases, lysozymes or muramidases, lytic transglycosylases, endopeptidases, and amidases. [169,170,171]. Gram-positive bacteria are more prone to endolysins, which are able to disrupt the peptidoglycan layer. Because gram-negative bacteria possess an outer membrane, shielding the peptidoglycan layer, the activity of endolysins towards gram-negative pathogens may be restricted [172] as the low permeability of the outer membrane makes the exogenous application of endolysins challenging [173] and some endolysins require additional strategies to permeate the outer membrane, such as endogenous lysis mediated by holins, pinholins, and bacterial Sec translocases or exogenous lysis [174]. Their medical potential is highlighted by the fact that, unlike for bacteriophages, there are no reports of bacterial resistance to endolysins acquired using classic mechanisms: by mutations, receptor modification, passive adaptation, restriction modification, CRISPR-Cas, and pseudolysogeny [175]. The most suitable and diverse endolysin described to date is Cpl-1 derived from S. pneumoniae bacteriophage Cp-1 [176]. Usually, bacteriophages are highly selective for a specific strain of bacteria; however, phage endolysins can display a broader activity against different and multiple hosts [166] and they can be easily engineered to enhance their lytic potential [177]. An increasing number of papers confirm their medical potential [178,179,180,181,182], but the results of clinical trials on phages have yet to prove the formal efficacy of PT.
除了完整的噬菌体病毒外，噬菌体内溶酶体最近也被证实在治疗严重的革兰氏阴性和革兰氏阳性细菌感染病例方面具有巨大潜力，如大肠杆菌、沙门氏菌属、铜绿假单胞菌、腐生假单胞菌、鲍氏志贺氏菌、柔性志贺氏菌、费氏弧菌、弧菌等[164, 165, 166, 167, 168]。根据肽聚糖破坏机制的不同，噬菌体内溶素又可细分为五个家族：葡糖胺酶、溶菌酶或呋喃酰胺酶、溶菌性转糖基酶、内肽酶和酰胺酶[169、170、171]。[ 169, 170, 171].革兰氏阳性细菌更容易感染内溶酶，因为内溶酶能够破坏肽聚糖层。由于革兰氏阴性细菌有一层外膜保护肽聚糖层，内溶素对革兰氏阴性病原体的活性可能会受到限制[172]，因为外膜的低渗透性使得外源性应用内溶素具有挑战性[173]，一些内溶素需要额外的策略才能渗透外膜，例如由 holins、pinholins 和细菌 Sec 易位酶介导的内源性裂解或外源性裂解[174]。与噬菌体不同的是，目前还没有关于细菌通过传统机制（突变、受体修饰、被动适应、限制性修饰、CRISPR-Cas 和假溶菌酶）对内生溶菌素产生抗药性的报道，这凸显了内生溶菌素的医学潜力[175]。迄今描述的最合适、最多样化的内溶素是源自肺炎双球菌噬菌体 Cp-1 的 Cpl-1[176]。通常情况下，噬菌体对特定菌株具有高度选择性；然而，噬菌体内溶酶体对不同和多种宿主具有更广泛的活性[166]，而且它们很容易被改造以增强其溶解潜能[177]。越来越多的论文证实了噬菌体的医疗潜力[ 178, 179, 180, 181, 182]，但噬菌体的临床试验结果尚未证明 PT 的正式疗效。