Molecular Epidemiology of Haemophilus influenzae Strains with Reduced Susceptibility and Genetic Profiles of Resistance in the Postvaccination Period

Abstract

Haemophilus influenzae serotype b has been the main cause of invasive infections in children, during the prevaccination period. More than 20 years after the introduction of the conjugate vaccine against Hib, HiNT has emerged as the cause of localized infections in children and adults. The main objective of this work is to evaluate the susceptibility and resistance mechanisms of H. influenzae strains from carriers and describe the molecular epidemiology and their clonal relationships by multilocus sequence typing (MLST). Sixty-nine strains from clinical cases and asymptomatic carriers from 2009 to 2019 were analyzed, confirmed as H. influenzae, and serotyped by polymerase chain reaction. The susceptibility to antibiotics was evaluated by E-test strips. Genotyping was performed by MLST. HiNT was the most frequent in all age groups. Resistance to ampicillin, sulfamethoxazole+trimethoprim, and amoxicillin+clavulanic acid was detected, with the production of β-lactamase being the main resistance mechanism. Among 21 HiNT strains with complete allelic MLST profiles, 19 new sequence types were described, reinforcing the already reported heterogeneity of nontypeable strains, and only one clonal complex (cc-1355) was observed. Our results show a high percentage of colonization regardless of age, increased antimicrobial resistance, and high genetic diversity, along with an increased number of cases caused by HiNT strains. These findings reinforce the need for continuous surveillance for HiNT strains as it has been reported worldwide after the introduction of the Hib conjugate vaccine.

Introduction

Haemophilus influenzae is a Gram-negative, immobile, nonspore-forming short rod-shaped pleomorphic bacterium.¹ H. influenzae strains may have a capsular polysaccharide, which is an important virulence factor, protecting the encapsulated bacteria against phagocytosis and opsonization.² Depending on the polysaccharide, there are six serotypes designated “a” through “f,” which are classified according to the antigenic characteristics of this structure. Strains that do not express the capsule are called nontypable (HiNT) or unencapsulated.³ Other virulence factors, in addition to the capsule observed in this species, is the presence of fimbriae, pili, adhesins, lipopolysaccharide, and bacteriocins.^2,4,5

H. influenzae may exist in the microbiota of the human upper respiratory tract mucosa as a natural host. However, progression to upper respiratory tract infections, pneumonia, or even sepsis and meningitis, among other infections, may occur.^6,7 In the prevaccination era, the vast majority of invasive diseases was caused by encapsulated serotype b strains (Hib). However, after the introduction of the conjugated vaccine, in some countries as in Brazil, there were drastic reductions of cases caused by this serotype and a reduction in the number of carriers among vaccinated and unvaccinated children.

At the same time, there was an increase in cases of invasive disease in adults, due to nonencapsulated strains in countries where the conjugate vaccine against Hib was introduced.^8,9 Thus, HiNT strains are currently the most frequent isolates of the species and the main cause of invasive diseases in all age groups, accounting for 62.5% of cases of H. influenzae diseases in children younger than 5 years (annual incidence of 1.73 per 100,000) and causing bacteremia in up to 75% of invasive infections.^3,10

In Brazil, the vaccination against Hib only started in 1999 with the conjugate vaccine polyribosylribitol phosphate (PRP) conjugated to tetanus toxoid for routine use in babies, as part of the national immunization program, which has been replaced by the combined diphtheria-tetanus-pertussis/Hib vaccine. Children <1 year of age were scheduled to receive three vaccine doses, given at 2-month intervals. The vaccination schedule followed in Brazil does not include the booster dose after 12 months. Children 12–23 months of age were scheduled to receive a single vaccine dose. Data from the Brazilian National Health Foundation/National Immunization Program (FUNASA/PNI/Ministry of Health) show that in 2002 and 2003, 92.91% and 95.99% of eligible children received the vaccine, respectively.¹¹

Since the introduction of Hib vaccine, the decrease of invasive Hib disease cases predominantly affecting children and the relative increase in nontypable H. influenzae infections have been well documented in many parts of the world. The increasing trend in the number of cases, especially among older adults, deserves attention.^10,12–15 However, there is a need of better understand the current scenario of H. influenzae infections in Brazil and its susceptibility profile, due to the lack of publications.¹⁶

Treatment and prophylaxis of diseases caused by H. influenzae are performed by antibiotic therapy, considering the sensitivity of the microorganism. β-Lactam antimicrobials and fluoroquinolones are frequently used as antimicrobial therapy for respiratory tract infections, showing good antibacterial activity against H. influenzae. However, increased resistance to ampicillin, amoxicillin+clavulanic acid, and second-generation cephalosporins has been reported. Today, the use of these drugs as a front line for the treatment of infections caused by this pathogen is problematic. Rifampicin is still the antimicrobial of choice for contact prophylaxis.^17–19

Several methods have been used to characterize H. influenzae strains, Multilocus Sequence Typing (MLST) being the most used, providing important data to characterize bacterial pathogen isolates.^8,20 The designation of Sequence Types (STs) and Clonal Complexes (cc) is possible to establish epidemiological relationships of the bacterial population.²¹

Although the implementation of the vaccine took place 20 years ago, only one study in Brazil used the MLST to assess the impact of the conjugate vaccine on the H. influenzae population. Strains from the city of Rio de Janeiro were analyzed and it was demonstrated that, although H. influenzae has become a relatively rare cause of invasive bacterial infections in this community after the introduction of the Hib conjugate vaccine, isolates from infectious cases became more diverse, with the emergence of HiNT and the increase of serotype a.¹⁶

Materials and Methods

The strains used in this study are part of the INCQS Bacterial Culture Collection/Hib vaccine sector. A total of 69 strains were selected, grouped as follows: 15 from vaccinated children (healthy carriers), collected from the population of a daycare center in Jacobina, Bahia, in 2011, and 54 clinical samples from infectious cases, collected from three states in Brazil, of which, 15 from São Paulo (2009–2016), 19 from Pernambuco (2009–2019), and 20 from Rio de Janeiro (2009–2019).

The biochemical characterization of H. influenzae was performed through carbohydrate fermentation (glucose, sucrose, lactose, and mannose), commercially supplied for preparation by Isofar, Merck, JT Baker, and Merck, respectively, and supplemented with hemin (factor X) and Nicotinamide Adenine Dinucleotide (NAD) (factor V), both from Sigma, and incubation at 35–37°C in a 5% CO₂ environment for 24 hours.²² Species confirmation was carried out later by polymerase chain reaction (PCR), with the amplification of the P6 gene and confirmation by sequencing.²³ Serotyping of H. influenzae strains was performed using the serum agglutination test (SAT) (Haemophilus influenzae Antisera; Difco) and subsequently confirmed by PCR, using specific primers for each capsular type targeting the bexA gene.²⁴ Strains that did not agglutinate for any of the six serotypes were initially identified as HiNT.⁵

Susceptibility to antimicrobials was carried out through an initial screening, by disk diffusion test in Haemophilus Test Medium (HTM) agar, which consists of Mueller Hinton base supplemented with 15 μg/mL of hemin, 15 μg/mL of NAD, and 5 mg/mL of yeast extract. For the disc diffusion test, the following antibiotics were used: ampicillin (10 μg), ceftriaxone (30 μg), chloramphenicol (30 μg), rifampicin (30 μg), sulfamethoxazole-trimethoprim (1.25/23.75 μg), clavulanate-amoxicillin (10/20 μg), levofloxacin (5 μg), and ciprofloxacin (05 μg).

After disc diffusion analysis, strains showing susceptibility profiles scored as R (resistant) or I (intermediate) were further tested by Minimum Inhibitory Concentration (MIC), using Etest^® strips, according to EUCAST,²⁵ and also by dilution method in HTM Agar. Only strains with decreased susceptibility to a specific antibiotic confirmed by MIC test were considered resistant. Strains with decreased resistance to β-lactam antibiotics were tested for β-lactamase production, using the chromogenic cephalosporin method, (Nitrocefin - Oxoid). For the susceptibility tests, we used the following control strains according to EUCAST: INCQS 01066 (ATCC 49247), for antibiotic assay quality control, and INCQS 01067 (ATCC 49766), for β-lactamase production assay quality control.

Capsular types were confirmed by PCR, according to Van Ketel et al. and Falla et al. Strains with positive amplification of P6 gene and failure to amplify bexA gene were classified as HiNT. The presence of genetic markers in beta-lactamase positive (BLP) strains was determined by PCR amplification targeting bla_TEM and bla_ROB-1 genes, as previously described.²⁶ β-Lactamase-negative strains with reduced susceptibility to β-lactam antibiotics were submitted to sequencing of the penicillin-binding protein (PBPs), for the presence of mutations in the ftsI gene.

MLST analysis was carried out with the objective to investigate the dynamics between the circulating strains of H. influenzae, in Brazil following the protocols described at (http://pubmlst.org/hinfluenzae/).⁸ Allele sequences obtained were compared with those deposited in the H. influenzae MLST database and allele numbers were assigned accordingly. Allele sequences that returned no match were submitted to the database for confirmation of new ST.

This study was approved by the CONEP ethics committee of Brazil under number 464174.

Results

Biochemical analysis of the 69 strains analyzed showed that the most common biotypes were II (36%), III (33%), and I (10%). Fourteen remaining strains were grouped into biotypes IV, V, VI, VII, VIII, and two strains not classified. Serotype analysis showed that 81% of the strains were classified as HiNT, which was also most frequent in all age groups. Serotypes a, b, and c were observed in 13 strains (19%), in relation to the total number of samples. Serotypes d, e, and f were not detected. Among 45 strains isolated from clinical samples, 34 (75.5%) were from invasive diseases, where 21 (61.7%) are HiNT and 13 (38.3%) belong to serotypes a (23.5%), b (8.8%), and c (5.8%). All strains isolated from noninvasive and unknown sources (n = 35), were classified as HiNT.

Capsulated strains (n = 13) of serotypes a, b, and c were all isolated from patients with invasive disease in all age groups, which makes 18.8% of all strains analyzed. Among patients with meningitis, 0 to 5 years of age (53%), a, b, and c serotypes showed a frequency of 29.4%, 17.6%, and 5.8%, respectively, within this group, and HiNT strains with a frequency of 47%. For other infections in this age group, HiNT strains were predominant (100%). HiNT strains were also predominant in age group older than 5 years, with 82.3%, while a and c showed frequencies of 11.7% and 5.9%, respectively. Table 1 shows the resistance profile of all strains analyzed.

Table 1.

Proportion of Haemophilus influenzae Strains Resistant to Different Antimicrobials

Antimicrobials	N (%)
AMP	17 (24.6)
SUT	9 (13)
AMX	3 (4.3)
AMP and SUT	6 (8.6)
AMX and AMP	1 (1.44)

AMP, ampicillin; AMX, amoxicillin +clavulanic acid; SUT, sulfamethoxazole+trimethoprim.

Excluding strains with no identified source of isolation, the proportion of invasive strains resistant to at least one antibiotic was lower (29.4%) than the percentage of resistant noninvasive strains (61.5%). While for ampicillin, MIC values of noninvasive strains ranged from 12 to >256 μg/mL, for invasive strains, MIC values were higher, ranging from 8 to >256 μg/mL. It is worth of note that 71.4% of noninvasive strains showed MIC values of >256 μg/mL for ampicillin. For sulfamethoxazole–trimethoprim, 100% of noninvasive resistant strains (n = 7) showed MIC values of >256/mL, while for the only two invasive resistant strains, MIC values were 1.5 and >256 μg/mL. For amoxicillin+clavulanic acid, invasive resistant strains showed higher MIC values ranging from 0.004 to >256 μg/mL when compared to noninvasive strains, which showed MIC values ranging from 0.5 to 8 μg/mL.

Only one invasive strain isolated from cerebrospinal fluid (CSF) of a 51-year-old patient showed an MIC value of >256 μg/mL for amoxicillin+clavulanic acid. This strain was highly resistant to ampicillin with MIC value >256 μg/mL as well. This strain was confirmed as a β-lactamase producer with a I449V mutation detected in the ftsI gene. From a total of 69 strains tested, 13 encapsulated strains with serotypes a, b, and c showed high antimicrobial susceptibility with only one strain resistant to chloramphenicol and two strains resistant to and amoxicillin+clavulanic acid with low MIC values (0.25 and 0.5 μg/mL). From a total of 56 HiNT strains, 37 (66%) were resistant to at least one antibiotic showing the highest MIC values for ampicillin, sulfamethoxazole+trimethoprim (SUT), and amoxicillin+clavulanic acid (Supplementary Table S1).

Sixteen strains resistant to ampicillin, were β-lactamase producers and classified as BLP (positive β-lactamase, ampicillin resistant). Three ampicillin-resistant strains, non β-lactamase producers, but showing mutations on the ftsI gene, were classified as β–lactamase-negative, ampicillin resistant (BLNAR). Gene bla_TEM was detected in all 16 BLP strains and bla_ROB gene was not found in any of the isolates. Three strains resistant to amoxicillin+clavulanic acid, non β-lactamase producers, were classified as β-lactamase-negative, amoxicillin+clavulanic acid resistant. Only one of these strains showed mutations in the ftsI gene. No type b^- or a^- strain was observed. Four SUT-resistant strains did not show β-lactamase production and were also classified as beta-lactamase negative, but ftsI gene was not sequenced for these strains (Table 2).

Table 2.

Minimum Inhibitory Concentration, Sequence Types, β-Lactamase Production, Minimum Inhibitory Concentration Values, and Amino Acid Substitutions of the ftsI Gene from 18 Samples

Strains	MIC (μg/mL)				Genes		Amino acid replacement (ftsI)
Strains	AMP	AMC	ST	β-lac	bla-_TEM	bla-_ROB	344 K	350 D	389 L	437 A	441 G	449 I	452 A	454 A	526 N^a
Jaci003	>256	-	2283^d	+	+	-	—	—	—	—	—	—	—	—	K
Jaci006	12	-	2284^d	+	+	-	—	—	F	—	—	—	—	—	—
Jaci010	>256	-	2285^d	+	+	-	—	—	—	—	—	—	—	—	K
Jaci013	16	-	2286^d	+	+	-	R	—	—	—	—	—	—	—	K
Jaci020	>256	-	2284^d	+	+	-	R	—	—	—	—	—	—	—	K
Jaci026	24	-	2284^d	+	+	-	—	—	—	—	—	—	—	—	K
Jaci031	16	-	2288^d	+	+	-	R	—	—	—	—	—	—	—	—
P3408	24	-	2281	+	+	-	R	—	—	—	—	—	—	—	K
P3932	-	6	2290	-	-	-	—	—	—	—	—	—	—	—	K
P4060	>256	-	2291	+	+	-	R	—	—	—	—	—	—	—	—
P4129	>256	-	ND	+	+	-	R	—	—	—	—	—	—	—	—
P4225	16	-	2294	-	-	-	—	—	—	—	—	—	—	—	K
P5522	32	-	2296^d	+	+	-	—	—	—	—	—	—	—	—	K
P5680^b	>256	-	ND	+	+	-	—	—	—	P	R	V	S	S	—
P6332^c	>256	>256	2297	+	+	-	—	—	—	—	—	V	—	—	—
P6337	8	-	2298	+	+	-	R	—	—	—	—	—	—	—	K
P6363	ND	-	—	-	-	-	—	N	—	S	—	—	—	—	K
P6364	ND	-	—	-	-	-	—	—	—	—	—	—	—	—	K
P6364	ND	-	—	-	-	-	—	—	—	—	—	—	—	—	K

Amino acid sequence of wild strain L42023.

Strain P5680 showed amino acid changes in three new positions not described in the literature: G441R, A452S, and A454S.

Strain P6332 showed high MIC values for both AMP and AMC.

A, alanine; AMC, amoxicillin+clavulanic acid; D, aspartate; F, phenylalanine; G, glycine; I, isoleucine; K, lysine; L, leucine; N, asparagine; ND, not determined; P, proline; R, arginine; S, serine; ST, Sequence Type; V, valine; β-lac, β-lactamase production.

Twenty-one strains were subjected to sequencing of the ftsI gene, which encodes the PBP3 protein, conferring resistance to β-lactams, to detect a possible relationship between the modification of the ftsI gene with β-lactam antibiotic resistance. Table 2 shows only the strains with alterations in the ftsI gene sequence. The DNA sequence of the ftsI gene of a wild strain sensitive to β-lactam antibiotics was used for comparison with the ftsI genes of the resistant strains. The sequence is deposited in GenBank under accession number L42023 (Haemophilus influenzae Rd KW20 chromosome) (Table 2).

Nineteen isolates classified as beta-lactamase positive amicacin resistant (BLPAR) were tested for the presence of β-lactamase-related genes bla_TEM and bla_ROB and sequencing of the ftsI gene was done to confirm the resistance mechanism present on those strains. Of these, three β-lactamase-negative strains showed alterations in the deduced amino acid sequences of the PBP3 protein. All three strains showed N526K mutation of the ftsI gene and one strain showed two additional changes in positions D350N and A437S (Table 2). Twenty-two out of the 23 strains submitted to molecular typing by MLST were resistant to at least one antibiotic. One strain susceptible to all antibiotics (Jaci015) was added for comparison with the resistant strains.

However, only 21 isolates with complete allelic profiles were obtained, since, in two resistant strains, it was not possible to determine the alleles of all constitutive genes. By MLST typing, only one strain matched to the previously defined ST-1355, first described in 2014 from a clinical case of meningitis in 2001 in Rio de Janeiro.¹⁶ Strain Jaci002, ST-1355, was isolated from an asymptomatic carrier in 2011, in Bahia. Nineteen new STs were described and submitted to the MLST database. Of the 21 strains with complete allelic profile, eight were also isolated from asymptomatic carriers in Bahia in 2011 (Table 3).

Table 3.

Multilocus Sequence Typing Results from Clinical and Carrier Strains

Isolates	adk	atpG	frdB	fucK	mdh	pgi	recA	ST	cc
P3408	1	141	1	70	15	53	38	2281	142
P3745	3	8	15	18	62	1	5	2282	836
Jaci002^a	44	2	16	108	17	2	3	1355	165
Jaci003^a	44	45	16	108	17	2	3	2283	165
Jaci006^a	44	2	194	108	17	2	3	2284	165
Jaci010^a	44	4	16	108	17	2	3	2285	165
Jaci013^a	44	7	16	108	17	2	3	2286	165
Jaci015^a	44	7	194	15	17	2	1	2287	98
Jaci020^a	44	2	194	108	17	2	3	2284	165
Jaci026^a	44	2	194	108	17	2	3	2284	165
Jaci031^a	44	2	175	108	239	2	3	2288	165
P3844	3	4	17	9	112	3	34	2289	166
P3932	10	22	15	30	1	21	1	2290	411
P4060	33	8	188	16	17	2	29	2291	107
P4084	28	33	19	19	58	17	2	2292	43
P4129	NA^b	2	16	14	9	14	130		ND^c
P4150	11	1	193	1	7	41	29	2293	597
P4225	28	33	19	14	15	1	2	2294	ND
P4257	11	2	7	14	28	26	3	2295	34
P5522	44	141	1	37	17	2	3	2296	165
P5680	NA	2	15	8	9	61	3		ND
P6332	14	6	16	5	26	61	3	2297	538
P6337	18	2	3	55	10	28	12	2298	18

New ST are in bold.

Strains isolated from healthy carriers.

No allele assigned.

No cc assigned.

cc, clonal complex.

Minimum Spanning Tree (MST) analysis in Fig. 1 shows the distribution of the 992 ST within all the 6,767 strains deposited in the MLST database so far. All MLST analyses were carried out based on the ST data available in the H. influenzae MLST database. All strains analyzed were classified with a new ST, except for one strain classified as ST-1355. Nineteen new STs (47.3%) were described and associated to 11 cc with nine STs associated to cc165, including ST-1355.

FIG. 1.

MST showing 6,605 strains and 2,620 ST deposited in the Haemophilus influenzae MLST database. The area where most of the strains analyzed in this study are allocated is indicated by a circle. MST, minimum spanning tree; MLST, multilocus sequence typing; ST, sequence types.

The remaining ST were associated to a single cc (Table 2). The area highlighted with a circle in Fig. 1 is showed with more details in Fig. 2 where the allocation of new ST, ST-1355, and cc165 is presented. The MST tree suggests that ST-503 is probably a common ancestor for the cc165 group formed through the following phylogenetic line: ST-6 (which is the ST with the larger number of strains) →ST-101 →ST-503 (cc165) → ST-1365 (cc165); at this point, there is a bifurcation leading to ST-165 and ST-1355 (cc165). Other ST not included in cc165 are scattered over the tree. Table 3 shows MLST alleles and designated new STs and cc.

FIG. 2.

MST with a zoom from Fig. 1, showing the allocation of ST-165 and related new ST included in cc165.

The MST analysis in Fig. 3 shows the distribution of ST among Brazilian isolates and worldwide ST-165 strains deposited in the H. influenzae MLST database. The strains are grouped by ST and serotypes. The group of ST belonging to cc165 is highlighted by a circle. It is worth of note that, despite there being nontypeable (NT) strains not included in cc165, all strains of this cc were characterized as NT and they were all derived from ST-6, which is only formed by serotype b strains.

FIG. 3.

MST of all Brazilian strains deposited in the Haemophilus influenzae MLST database comparing ST and serotypes. Strains included in cc165 are indicated by a circle.

When looking at data provided by the MLST database, ST-6 strains (n = 327) are classified as serotype b (98.2%), serotype a (0.3%), and NT (1.5%) strains. The only two strains of this group isolated in Brazil were from a case of co-infection with b and b^- strains causing bacteremia and meningitis in the same patient.²⁷ Two strains of this group were from carriage and 86.5% of the strains where isolated in Europe. When we look at the world distribution of STs and serotypes (Fig. 1), ST-6 appears to be a great producer of NT strains, which is confirmed with the distribution of STs and serotypes in Brazil, as shown in Fig. 3.

Figure 4 shows the distribution of the same group of strains, comparing ST with the origin (disease or carriage). Again, cc165 is strongly related to carriage, with the only exception of ST-2296, which was isolated from a patient with bacteremia and shares only four alleles with ST-165, while all the other carriage strains within cc165 share 5 or more alleles. Only a few other strains not included in cc165 were also isolated from carriers, as well as all the close-related cc165 strains.

FIG. 4.

MST with strains isolated in Brazil deposited in the Haemophilus influenzae MLST database comparing ST and source of isolation (clinical or carrier). Strains included in cc165 are indicated by a circle.

Discussion

It is estimated that the use of the conjugate vaccine against Hib, implemented in the childhood vaccination calendar in 1999, has prevented the serious neurological sequelae found in more than 25% of Hib meningitis survivors during the prevaccination period.⁹ Currently, HiNT is the most frequent serotype isolated among H. influenzae strains and is the main cause of invasive diseases in all age groups.^3,10 Considering this information, there is a need to better understand the current scenario of H. influenzae infections in Brazil.

As noted, in this study, the percentage of colonization was higher among healthy children than among adults; although the rate of colonization in adults by H. influenzae is growing, it is still lower than that observed in children, even in the postvaccination era. This corroborates what has been reported in other countries such as Belgium and Spain that reported high rates of colonization in children by H. influenzae even in the postvaccination era.^28,29 However, when comparing the percentages of colonization by serotype, both in adults and in children, HiNT is the most frequent in relation to the encapsulated strains and the total number of strains for each age group studied.

Likewise, according to Yang et al., NT strains presented a percentage of 22% of colonization among healthy children in China.³⁰ Thus, we observed that HiNT were equally capable of producing clinical cases among adults, confirming that this type of H. influenzae has become increasingly virulent, as described in the literature, in any age group, which classifies it as an emerging concern among these strains in Brazil and worldwide. However, although NT strains were more frequent in all age groups, there was no close relationship of these strains with cases of invasive disease in patients younger than/equal to 5 years, and encapsulated strains were more frequent compared to NT, as it has been reported in Japan in 2018.³¹

Although the Hib vaccine has reduced the incidence of cases caused by this serotype, in our study, three out of nine encapsulated strains that caused invasive disease in children younger than 5 years were Hib. This shows that, although low colonization by serotype b is still present, confirming what previous studies already described, the effectiveness of the vaccine even 20 years after its implementation is unquestionable.

The results of the susceptibility assays (agar dilution and MIC) with antibiotics, ceftriaxone, chloramphenicol, levofloxacin, ciprofloxacin, and rifampicin, showed high in vitro activity against H. influenzae, and therefore may serve as the basis for the therapeutic choice of infections caused by this pathogen. However, high resistance to ampicillin and sulfamethoxazole+trimethoprim was observed. Although only five strains (7.2%) showed resistance to amoxicillin+clavulanic acid, a reduction in susceptibility to this antimicrobial in Brazilian strains is noted. The reduction in susceptibility to β-lactams such as ampicillin has been reported for years. In this study, one strain showed high resistance to this antimicrobial with MIC level >256 μg/mL.

The main mechanism of resistance to β-lactam antimicrobials within the strains analyzed is the production of β-lactamase enzyme, mediated by the bla_TEM gene, since the bla_ROB gene was not found in any of the BLP strains. These findings corroborate those reported by Farrell et al., in 2005, with H. influenzae isolates from several countries, including Brazil, where they found an average of 93.7% positive strains for the bla_TEM gene.³² Another known mechanism involved in β-lactam antibiotic resistance, linked to increased resistance in BLNAR, is nucleotide changes of the ftsI gene encoding the PBP3 protein. In this work, we identified alterations in 17 β-lactamase-producing strains, with at least one mutation in their amino acid sequences, such as N526K, which was observed as the most frequent among BLP strains (64.7%), which has been also reported in other studies.^33,34

Previous studies have reported a reduction in the susceptibility of H. influenzae strains to amoxicillin+clavulanic acid, and the prevalence of BLNAR and BLPAR strains harboring mutations in ftsI and production of β-lactamase, which is of growing concern. In our study, we report a strain classified as BLPAR with a MIC >256 μg/mL for both amoxicillin+clavulanic acid belonging to ST-2297, which is not associated to cc165. All strains included in cc165, however, were classified as β-lactamase producers harboring the bla-_TEM gene and carry mutations in the ftsI gene coding for an altered PBP3. All PBP3 mutations reported in this study have been already described elsewhere and are associated with decreased susceptibility to β-lactam antibiotics, including the N526K (Ubukata group II)³⁵ substitution, which is frequently associated with BLNAR strains.^33,36

The S385T substitution, which is common in resistant strains circulating in Japan and South Korea, was not found in our strains. However, three new mutations were reported on a strain with high MIC value for ampicillin and β-lactamase producer, G441R, A452S, and A454S. Such data may suggest a tendency among Brazilian strains to develop both resistance mechanisms, which make the genetic analysis of ftsI mutations in both BLP and BLNAR strains of great importance for epidemiological surveillance.

All 21 isolates with a complete MLST profile were characterized as NT. Nine (42.8%) strains were grouped into cc165, one (4.7%) designated as ST-2294 was not included in a cc, and 11 (52.3%) sharing 4 or more alleles to a previously described ST were associated to that cc. A total of 20 new STs were assigned. The only ST identified in this study matching with an ST previously deposited in the database was ST-1355, described and identified for the first time in 2017.¹⁶ These data suggest a high genetic diversity and an ongoing evolution of H. influenzae strains in Brazil among carrier and clinical strains, corroborating what has already been described about heterogeneity of HiNT strains.¹⁰

Strain P6332 (ST-2297) described in this study shares four alleles with ST-538 (frdB-16, mdH-26, pgI-61, and recA-3). This strain was classified as BLPAR with an MIC >256 μg/mL and ftsI amino acid replacement I448V, isolated from a carrier in 2018. Only one strain belonging to ST-538 has been described from a carrier (female, 4-year-old child) in South Korea in 2008. Other 58 strains sharing four or more alleles with ST-538 were described from different countries of the 5 continents. The emergence of a carrier strain with such level of restistance to a β-lactam antimicrobial, should be a real concern to health surveillance authorities since the spread of these strains is a real threat.

The MST analysis of isolates from Brazil, shown in Fig. 3, suggests a strong association with HiNT strains associated to cc165, which was the prevalent cc in this study. Worldwide, 75 strains from several STs have been associated with cc165 sharing four or more alleles with ST-165. Within this group, 10 strains (13.3%) were isolated in Brazil. All strains belonging to cc165 are classified as HiNT and were isolated from all five continents, from 1998 to 2020, and several different sources (carrier, invasive disease, otitis, pneumonia, etc.).

Other cc where Brazilian strains were allocated have a major representation in the whole database, such as cc107 with 227 strains (ST-2291), cc34 with 120 strains (ST-2295), cc18 with 226 strains (ST-2298), and cc98 with 220 strains (ST-2287), among others. In these groups the representation of Brazilian isolates is much lower, cc107 (0.4%), cc34 (1.6%), cc18 (0.9%), and cc98 (0.4%). This finding suggests that cc165, with a large number of Brazilian strains of different ST, is the most common cc in Brazil and should be closely monitored since it is formed by a cluster of closely related strains with invasive profiles to antibiotics.

The expansion of HiNT from ST-6 formed almost exclusively by Hib strains (96%) is evident worldwide and much more among the Brazilian isolates that appear with only 1.4% of the isolates belonging to cc6, confirming what has already been reported by previous studies.¹⁰ These data once again show the expansion of HiNT within isolates of this species and consequently, its clinical importance and need for control. The expansion of NT strains from cc6, formed exclusively by serotype b strains, shows the adaptive potential of this species, which by changing the biochemistry of its capsule facilitates the escape of the immune system of vaccinated individuals.^37–39

It is important to note that, according to the MLST database, strains belonging to ST-6 were first isolated in 1946 and 1950^8,40 with two serotype b strains from United States and has been persistently isolated so far, with the most recent Hib strains isolated in 2020 in France and 2022 in the Czech Republic. It is also worth of note that within the group of ST-6 strains, two Hib and Hib^- strains are included, isolated in Brazil in 2009 from blood and CSF from the same patient with septicemia and meningitis, which may be additional evidence of the differentiation process of serotype b strains that apparently became NT.²⁷

Footnotes

Acknowledgments

We thank the Central Reference Laboratories (LACEN-PE, LACENRJ, and LACEN-BA) and Hospital Municipal Lourenço Jorge-RJ for the clinical samples supplied. This publication made use of the Multi Locus Sequence Typing website () developed by Keith Jolley and sited at the University of Oxford. The authors are grateful to the Sequencing core “PlataformaGenômica de Sequenciamento de DNA/PDTIS-FIOCRUZ.”

Authors' Contributions

I.F. and A.E.C.C.A.: conceptualization, I.M.B., A.E.C.C.A., and I.F.: data curation and formal analysis, I.F. and A.E.C.C.A.: funding acquisition, I.M.B., C.F.A., D.M.O., N.G.S.C., J.R.O., and A.P.A.N.: investigation and methodology, I.M.B., I.F., and A.E.C.C.A.: project administration resources and supervision, I.M.B., C.F.A., and A.P.A.N.: validation, I.M.B. and I.F.: visualization, IMB, I.F., and A.E.C.C.A.: writing original draft, review, and editing.

Disclosure Statement

No competing financial interests exist.

Funding Information

Programa de Pós-Graduação em Vigilância Sanitária—PPGVS (INCQS/FIOCRUZ), Consellho Nacional de Desenvolvimento e Pesquisa (CNPq), and Fundação de Amparo à Pesquisa do Estado do Rio de Janeiro (FAPERJ), INCQS/FIOCRUZ and Coordenação de Aperfeiçoamento de Pessoal de Nível Superior—Brasil (CAPES)—FinanceCode 001. The development of the PUBMlst website () has been funded by the Welcome Trust and European Union.

Supplementary Material

References

Marrs

, Krasan

, McCrea

, et al. Haemophilus influenzae—Human specific bacteria. Front Biosci, 2001; 6(4):41–60.

Madigan

. ed. Microbiologia de Brock. Printice Hall, Artmed: São Paulo; 2016.

Butler

, Myers

. Changing epidemiology of Haemophilus influenzae in children. Infect Dis Clin North Am, 2018; 32(1):119–128.

Häuser

, Wegele

, Stump-Guthier

, et al. Capsule and fimbriae modulate the invasion of Haemophilus influenzae in a human blood-cerebrospinal fluid barrier model. Int J Med Microbiol, 2018; 308(7):829–839.

Koneman

, Allen

, Janda

, et al. eds. Koneman's Color Atlas and Text of Diagnostic Microbiology, 5 ed. Editora Médica e Científica: Rio de Janeiro; 2001.

Bogaert

, Keijser

, Huse

, et al. Variability and diversity of nasopharyngeal microbiota in children: A metagenomic analysis. PLoS One, 2011; 6(2):e17035.

Kosikowska

, Korona-Głowniak

, Niedzielski

, et al. Nasopharyngeal and adenoid colonization by Haemophilus influenzae and Haemophilus parainfluenzae in children undergoing adenoidectomy and the ability of bacterial isolates to biofilm production. Medicine, 2015; 94(18):e799.

Meats

, Feil

, Stringer

, et al. Characterization of encapsulated and noncapsulated Haemophilus influenzae and determination of phylogenetic relationships by multilocus sequence typing. J Clin Microbiol, 2003; 41(4):1623–1636.

Ribeiro

, Lima

, Reis

, et al. Haemophilus influenzae meningitis 5 years after introduction of the Haemophilus influenzae type b conjugate vaccine in Brazil. Vaccine, 2007; 25(22):4420–4428.

10.

Langereis

, de Jonge

. Invasive disease caused by nontypeable Haemophilus influenzae. Emerg Infect Dis, 2015; 21(10):1711–1718.

11.

FUNASA—Boletim Eletrônico Epidemiológico. Available at: www.funasa.gov.br (Last accessed: June 12, 2003).

12.

Heath

, Booy

, Azzopardi

, et al. Non-type b Haemophilus influenzae disease: Clinical and epidemiologic characteristics in the Haemophilus influenzae type b vaccine era. Pediatr Infect Dis, 2001; 20(3):300–305.

13.

De Almeida AECC, de Filippis

, O de Abreu

, et al. Occurrence of Haemophilus influenzae strains in three Brazilian states since the introduction of a conjugate Haemophilus influenzae type b vaccine. Braz J Med Biol Res, 2005; 8(5):777–781.

14.

Dworkin

, Park

, Borchardt

. The changing epidemiology of invasive Haemophilus influenzae disease, especially in persons > or = 65 years old. Clin Infect Dis, 2007; 44(6):810–816.

15.

TSANG

RSW

, et al. Characterization of invasive Haemophilus influenzae disease in Manitoba, Canadá, 2000–2006: Invasive disease due to non-type B strains. Clin Infect Dis, 2007; 44(12):1611–1614.

16.

Tuyama

, Corrêa-Antônio

, Schlackman

, et al. Invasive Haemophilus influenza disease in the vaccine era in Rio de Janeiro, Brazil. Mem Inst Oswaldo Cruz, 2017; 112(3):196–202.

17.

Cherkaoui

, Gaïa

, Baud

, et al. Molecular characterization of fluoroquinolones, macrolides, and imipenem resistance in Haemophilus influenzae: Analysis of the mutations in QRDRs and assessment of the extent of the AcrAB-TolC-mediated resistance. Eur J Clin Microbiol Infect Dis, 2018; 37(11):2201–2210.

18.

Eliakim-Raz

, Lador

, Leibovici-Weissman

, et al. Efficacy and safety of chloramphenicol: Joining the revival of old antibiotics? Systematic review and meta-analysis of randomized controlled trials. J Antimicrob Chemother, 2015; 70(4):979–996.

19.

Lopes

(ed). Tratado de Clínica Médica. Roca: São Paulo; 2006.

20.

Jolley

, Bray

, Maiden

. Open-access bacterial population genomics: BIGSdb software, the PubMLST.org website and their applications. Wellcome Open Res, 2018; 3:124.

21.

Maiden

, Bygraves

, Feil

, et al. Multilocus sequence typing: A portable approach to the identification of clones within populations of pathogenic microorganisms. Proc Natl Acad Sci U S A, 1998; 95(6):3140–3145.

22.

Ledeboer

, Doern

. Haemophilus. In: Manual of Clinical Microbiology 10th ed. (Versalovic J. ed). American Society for Microbiology: Washington, DC, USA; 2011; Cap. 34, pp. 588–602.

23.

Van Ketel

, de Wever

, van Alphen

. Detection of Haemophilus influenzae in cerebrospinal fluids by polymerase chain reaction DNA amplification. J Med Microbiol, 1990; 33(4):271–276.

24.

Falla

, Crook

, Brophy et al. PCR for capsular typing of Haemophilus influenzae. J Clin Microbiol, 1994; 32(10):2382–2386.

25.

European Committee on Antimicrobial Susceptibility Testing. The European Committee on Antimicrobial Susceptibility Testing—EUCAST. 2019. Available at: https://eucast.org/fileadmin/src/media/PDFs/EUCAST_files/Breakpoint_tables/v_9.0_Breakpoint_Tables.pdf (Last accessed: April 11, 2020).

26.

García-Cobos

, Campos

, Cercenado

, et al. Antibiotic resistance in Haemophilus influenzae decreased, except for beta-lactamase-negative amoxicillin-resistant isolates, in parallel with community antibiotic consumption in Spain from 1997 to 2007. Antimicrob Agents Chemother, 2008; 52(8):2760–2766.

27.

Batalha

, de Almeida

, Caldeira

, et al. Changes in Haemophilus influenzae capsule locus: Possible emergence of novel variants in Brazil. Diagn Microbiol Infect Dis, 2010; 68(2):97–102.

28.

Jourdain

, Smeesters

, Denis

, et al. Differences in nasopharyngeal bacterial carriage in preschool children from different socio-economic origins. Clin Microbiol Infect, 2011; 17(6):907–914.

29.

Puig

, Grau

, Marti

, et al. Clinical and molecular epidemiology of Haemophilus influenzae causing invasive disease in adult patients. PLoS One, 2014; 9(11):e112711.

30.

Yang

, Zhang

, Peng

. The pharyngeal carriage of Haemophilus influenzae among healthy population in China: A systematic review and meta-analysis. BMC Infect Dis, 2019; 19(1):547.

31.

Naito

, Takeuchi

, Ohkusu

, et al. Clinical and bacteriologic analysis of nontypeable Haemophilus influenzae strains isolated from children with invasive diseases in Japan from 2008 to 2015. J Clin Microbiol, 2018; 56(7):e00141–18.

32.

Farrell

D J

, Morrissey

, Bakker

, et al. Global distribution of TEM-1 and ROB-1 beta-lactamases in Haemophilus influenzae. J Antimicrob Chemother, 2005; 56(4):773–776.

33.

Skaare

, Anthonisen

, Caugant

, et al. Multilocus sequence typing and ftsI sequencing: A powerful tool for surveillance of penicillin-binding protein 3-mediated beta-lactam resistance in nontypeable Haemophilus influenzae. BMC Microbiol, 2014; 14:131.

34.

Shuel

, Tsang

. Canadian beta-lactamase-negative Haemophilus influenzae isolates showing decreased susceptibility toward ampicillin have significant penicillin binding protein 3 mutations. Diagn Microbiol Infect Dis, 2009; 63(4):379–383.

35.

Ubukata

, Shibasaki

, Yamamoto

, et al. Association of amino acid substitutions in penicillin-binding protein 3 with beta-lactam resistance in beta-lactamase-negative ampicillin-resistant Haemophilus influenzae. Antimicrob Agents Chemother, 2001; 45(6):1693–1699.

36.

Søndergaard

, Witherden

, Nørskov-Lauritsen

, et al. Interspecies transfer of the penicillin-binding protein 3-encoding gene ftsI between Haemophilus influenzae and Haemophilus haemolyticus can confer reduced susceptibility to β-lactam antimicrobial agents. Antimicrob Agents Chemother, 2015; 59(7):4339–4342.

37.

Mironov

, Tagachenkova

, Platonov

, et al. Genetic characteristic of Haemophilus influenzae type B strains isolated in Russian regions. Zh Mikrobiol Epidemiol Immunobiol, 2010;(4):3–8.

38.

Krízová

, Lebedová

, Benes

. The effect of routine vaccination in the Czech Republic on the incidence of invasive diseases caused by Haemophilus influenzae b. Klin Mikrobiol Infekc Lek, 2004; 10(3):118–123.

39.

Krízová

, Kalmusová

, Lebedová

, et al. Characterization of Haemophilus influenzae strains using multilocus sequencing. Epidemiol Mikrobiol Imunol, 2004; 53(2):74–77.

40.

Hewitt

, Pittman

. Antibacterial action of penicillin, penicillin X, and streptomycin on Hemophilus influenzae. Public Health Rep, 1946; 61:768–778.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.15 MB